multi-dimension model of memory in ocd: a-meta analysis

RUNNING HEAD: MULTI-DIMENSION MODEL OF MEMORY IN OCD: A-META ANALYSIS

Title

Modelling a Multi-Dimensional Model of Memory Performance in Obsessive-Compulsive Disorder: A Multi-Level Meta-Analytic Review.

Authors

Persson, Sc., Yates, Aa., Kessler, Kb., & Harkin, Ba*

a Department of Psychology, Manchester Metropolitan University.

b Aston Neuroscience Institute, Aston University.

c School of Social Sciences, Leeds Beckett University

* Corresponding author: Dr Ben Harkin, Department of Psychology, Manchester Metropolitan University, All Saints Building, Manchester, M15 6BH. E-mail: [email protected]

KK was supported by MRC grant MR/J001953/1

1

mailto:[email protected]


Abstract

Even though memory performance is a commonly researched aspect of Obsessive-Compulsive

Disorder (OCD), a coherent and unified explanation of the role of specific cognitive factors has

remained elusive. To address this, the present meta-analysis examined the predictive validity of

Harkin and Kessler’s (2011) Executive Function (E), Binding Complexity (B) and Memory Load (L)

Classification System with regards to affected vs. unaffected memory performance in OCD. We

employed a multi-level meta-analytic approach (Viechtbauer, 2010) to accommodate the

interdependent nature of the EBL model and interdependency of effect sizes (305 effect sizes from

144 studies, including 4424 OCD patients). Results revealed that the EBL model predicted memory

performance, i.e., as EBL demand increases, those with OCD performed progressively worse on

memory tasks. Executive function was the driving mechanism behind the EBL’s impact on OCD

memory performance and negated effect size differences between visual and verbal tasks in those

with OCD. Comparisons of sub-task effect sizes were also generally in accord with the cognitive

parameters of the EBL taxonomy. We conclude that standardised coding of tasks along individual

cognitive dimensions and multi-level meta-analyses provides a new approach to examine multi-

dimensional models of memory and cognitive performance in OCD and other disorders.

Keywords

Obsessive-Compulsive Disorder; Memory Performance; Executive Function; Binding Complexity,

Memory Load, Multi-Level Meta-Analysis.

2


Even though memory performance is a commonly researched aspect of Obsessive-

Compulsive Disorder (OCD), a coherent and unified explanation of the role of specific cognitive

factors has remained elusive (Hermans, Engelen, Grouwels, Joos, Lemmens, & Pieters, 2008; Snyder,

Kaiser, Warren, & Heller, 2015). Historically, there has been a tendency in the literature to view

verbal and visual memory/tasks as distinct entities (e.g., Boone, Ananth, Philpott, Kaur, &

Djenderedjian, 1991; Zielinski, Taylor, & Juzwin, 1991; Christensen, Kim, Dysken, & Hoover, 1992;

Muller & Roberts, 2005) and to focus on general (e.g. long-term) mnestic performance (McNally &

Kohlbeck, 1993; MacDonald, Antony, Macleod, & Richter, 1997; Tallis, 1997; Jelinek, Moritz,

Heeren, & Naber, 2006). An alternative perspective indicated a more subtle relationship, with

memory impairment secondary to executive dysfunction (Greisberg & McKay, 2003), wherein

deficits in executive function in conjunction with task demands differentiate the memory performance

of those with OCD from controls (Olley, Malhi, & Sachdev, 2007). Extending the latter, Harkin and

Kessler (2011) proposed a tripartite explanation of memory impairments in OCD, i.e., they occur

when a task taps into specific aspects of executive dysfunction (E), depends upon binding and/or

chunking of complex information (B) and/or places a significant load on memory capacity (L) (Figure

1; Harkin & Kessler, 2011a). In effect, the EBL (executive functioning, binding complexity, memory

load) classification system provided a qualitative explanation of disparate memory findings. The

present meta-analysis provides the next logical step, in that it aims to standardise dimensions of the

EBL taxonomy and then quantify how they moderate memory performance in OCD.

The Executive-Functioning, Binding Complexity, Memory Load (EBL) Classification System

The original catalyst for the EBL classification system was the growing body of research that

indicated memory impairments were secondary to executive dysfunction, with general memory

capacity remaining intact (Olley, et al., 2007; Omori, Murata, Yamanishi, Nakaaki, Akechi, Mikuni,

& Furukawa, 2007; Cha, Koo, Kim, Kim, Oh, Suh, & Lee, 2008; Exner, Martin, & Rief, 2009).

Specifically, in a series of delayed-match to sample working memory (WM) experiments, Harkin et

al. presented a range of stimuli (e.g., letters in locations; kitchen appliances on a stove) to be

remembered over a short delay (see Harkin & Kessler, 2009; Harkin, Rutherford, & Kessler, 2011;

Harkin & Kessler, 2011b; Harkin, Miellet, & Kessler, 2012b). Memory impairment for subclinical

OCD-checkers only occurred when misleading and irrelevant information (e.g., asking the location of

a letter or kitchen appliance that was not part of the original encoding set) was presented between the

encoding set and the memory probe. In agreement with other research (eg., van der Wee, Ramsey,

Jansma, Denys, van Megen, Westenberg, & Kahn, 2003; Ciesielski, Hamalainen, Geller, Wilhelm,

Goldsmith, & Ahlfors, 2007; Henseler, Gruber, Kraft, Krick, Reith, & Falkai, 2008) an impairment in

general WM capacity could not explain these findings, as across various iterations of the basic

paradigm, performance was intact in the absence of a misleading or irrelevant distractor (see Harkin

3

LOWBINDING

REQUIREMENT(UNIMODAL)

(B)

LOWLOAD

(L)

HIGHBINDING

REQUIREMENT(MULTIMODAL)

(B)

HIGHLOAD

(L)

LOWEXECUTIVE

DEMAND(E)

HIGHEXECUTIVE

DEMAND(E)

KEY: LOCATION AND SEVERITY OF OCD MEMORY DEFICIT COMPARED TO CONTROLS

Increased chance for OCD-specific Memory Deficit

No OCD-specific Memory Deficit compared to Controls


& Kessler, 2009; Harkin, et al., 2011; Harkin & Kessler, 2011b; Harkin, et al., 2012b). To explain the

findings, Harkin and Kessler (2009) drew upon Baddeley’s updated model of WM (Baddeley, 2000),

wherein, in addition to the phonological loop and visuospatial sketchpad of the original model

(Baddeley, 1986), Baddeley (2000) included an “episodic buffer” to explain the integration of

temporary, multimodal representations in WM. This provided a solution to the “binding problem”

(Treisman, 1996), as in reality stimuli are rarely presented in isolation, but rather are embedded as a

multi-featured object (size, shape, colour, semantics, etc.), in a location, within a complex scene and

context (Hinton, McClelland, & Rumelhart, 1986). The binding and maintenance of these “fragile”

multimodal representations occurs via the central executive, which explains the WM performance in

the tasks of Harkin and colleagues (Allen, Baddeley, & Hitch, 2006). As such, we proposed (Harkin

& Kessler, 2009) that an executive dysfunction (e.g., unsuppressed intrusive thoughts/stimuli) in

those with OCD interfered with fragile multimodal bindings in the EB (i.e., letters/electrical

appliances to locations), which impaired the consolidation of affected episodes into WM and long-

term memory (LTM).

Figure 1. The EBL Classification System, adapted from Harkin and Kessler (2011a). It is important to note that Harkin and Kessler’s three EBL dimensions are not conceived of as fully orthogonal. Binding complexity might affect memory load in circumstances where large numbers of features needto be bound. Complex bindings as well as increased load will draw on executive functions when exceeding limitations, thus, executive demands are proposed as the most fundamental dimension (Harkin & Kessler, 2011).

A Review of Reviews on Memory Performance in OCD

4


Many excellent reviews and meta-analyses have contributed to how we understand memory

performance in OCD. We will show that they identify the parameters – either implicitly or explicitly

– of memory performance in OCD as expected based on Harkin and Kessler’s (2011a) EBL

taxonomy.

Systematic reviews: Visual versus verbal memory performance in OCD. Greisberg and

McKay (2003) reviewed findings on attention, executive function, and memory in OCD. They

reported that memory impairments (visual and verbal) in OCD are not attributable to issues of basic

capacity per se but rather the organizational demands of the task. They then went on to explain how

task demands (low vs. high) explained memory performance (absent vs. present, respectively) in

OCD. Related to low overall EBL task demands, they stated that “when . . . tasks [demand] recall

under well-structured circumstances, those with OCD . . . perform . . . similar to those without OCD"

(Greisberg & McKay, 2003, p.110). Related to memory performance at high overall EBL task

demands, they stated that “when tasks are less clearly defined, or when the ability to recall

information . . . [requires] a combination of memory and organization . . . (as in the RCFT), then

significant impairment becomes evident” (Greisberg & McKay, 2003, p.110). From this, we infer that

it is task complexity (i.e. load, bindings, executive demands), which determines affected vs.

unaffected memory performance in OCD.

Kuelz, Hohagen, and Voderholzer (2004) reported a varied pattern of results. To begin with,

basic capacity was generally intact (e.g., WAIS-R Digit Span forward; M. D. Lezak, 1995), but the

authors then reported a diverse pattern of findings for verbal/non-verbal fluency and higher-order

executive functions like planning ability (e.g., Tower of London: TOL; Shallice, 1982). They also

reported specific and consistent impairments on complex visuospatial reproduction tasks (e.g.,

RCFT). They proposed that memory impairments were “secondary to an inability to apply efficiently

elaborated strategies” (p. 209). Wherein, those with OCD focus on irrelevant details during the

encoding and copy phases of such tasks (see Harkin, et al., 2012b). Three selective reviews by Muller

and Roberts (2005), Olley et al (2007), Abramovitch, and Cooperman (2015) further underlined these

conclusions. With inconsistent results for the recall and recognition of verbal information, and

reliable deficits in the memory of complex visual material. In sum, they attributed this to visual tasks

(i.e., high EBL demand) exposing the inabilities of those with OCD to generate and implement

organizational strategies (E) to encode complex visuospatial patterns (B and L).

Meta-analytic reviews: The importance of executive dysfunction. As the number of meta-

analysis increased, a more nuanced pattern of findings emerged. For example, in a meta-analysis of

113 studies, Abramovitch, Abramowitz, and Mittelman (2013) examined various cognitive domains

(e.g., attention, executive functions, visuospatial abilities, WM) and reported the classic large versus

small effect size for visual (d=-0.76) and verbal memory (d=-0.33) and medium effect sizes for a

range of executive tests. They also reported that executive dysfunctions (e.g., set-shifting) were only

5


associated with impairments in visual but not verbal memory and that impairment in visual memory

“may be related to executive functioning and less with memory impairment per se” (p. 1168). A more

complex pattern of memory impairments was then reported in a meta-analysis by Shin, Lee, Kim, and

Kwon (2014). From 88 studies, they reported that those with OCD were impaired in a range of

cognitive tasks across executive, verbal and visual domains. For example, the largest impairments

were again observed on visual tasks like the RCFT (g=-0.74), TOL (g=-0.73) and executive

organization (g=-0.63); medium effects for verbal tasks (e.g., verbal learning memory-II; g=-0.42);

and no significant impairments on the digit span task (g=-0.11). From this, we infer that memory

impairments are most likely due to the extent that tasks tap into the executive and/or organizational

abilities of OCD participants, as opposed to a simple dissociation between visual or verbal tasks.

Similarly, in a meta-analysis of 101 studies, Snyder et al. (2015) reported that those with

OCD suffered from a global impairment (i.e., d=0.3-0.5) across a range of executive tasks (i.e.,

inhibition, shifting, updating, verbal fluency, planning, general motor speed, and verbal/visuospatial

WM). First, ‘updating’ had the largest overall effect size of d=0.71 for the n-back task, whereas,

within the ‘visuospatial WM’ category, effect sizes for self-ordered pointing, composite score and

block span were d=0.62, 0.47 and 0.43, respectively. Second, within the ‘verbal WM’ domain,

manipulation of verbal information had a small but significant effect (d=0.31), whereas simple

maintenance (d=0.07) and digit span forward (d=0.08) had very small and non-significant effects. A

similar pattern was observed in a meta-analysis conducted by Leopold and Backenstrass (2015), who

reported that the largest impairments were observed in tests of sustained attention, encoding, verbal

and visual memory. This led them to propose a link between “applying organizational strategies to

the encoding of verbal and nonverbal [emphasis added] information … [and] poorer memory

performance in OCD patients” (p. 56).

Collectively, these reviews and meta-analyses highlight the following key points. (a) Memory

impairment in OCD is secondary to executive dysfunction. (b) The visual versus verbal distinction

might be of secondary importance to the underlying demands of the memory task. (c) Irrespective of

domain (visual or verbal) memory impairment in OCD is likely when the tasks require a high degree

of executive control (i.e., organizational strategies, chunking, updating, sorting) upon the task-related

contents maintained in WM. (d) Memory impairment in OCD is likely when tasks are high in binding

complexity (B) and/or load (L) and require organizational strategies (E) in service of such task

demands. (e) There is a need to examine memory performance at a domain/sub-task level, as

averaging across these will obscure unique contributions of different EBL demands to memory

performance in OCD.

Theoretical and Empirical Foundations of the EBL Classification System

6


The previous discussion suggests that the EBL taxonomy offers a parsimonious means to

explain, classify and predict the often-complex pattern of memory impairments that are observed in

OCD. It is important to note and as we state in Figure 1 that the dimensions are interdependent (i.e.,

non-orthogonal) as originally conceived in our 2011 paper (Harkin & Kessler, 2011a). In that,

binding complexity might affect memory load in circumstances where large numbers of features need

to be bound. Complex bindings as well as increased load will draw on executive functions when

exceeding limitations, thus, executive demands are proposed as the most fundamental dimension of

our taxonomy. We now detail each dimension of the EBL and highlight the theoretical and empirical

foundations to each:

(1) Executive function (E). We adopt Walter and Raffone’s (2008) tripartite explanation of

executive functioning of (a) Attentional Control: top-down selective activation of task-

relevant representations and inhibition of task-irrelevant stimuli and responses (see also

Adele, 2013); (b) Maintenance and Updating: focus on and hold task-relevant information in

an active state, and when required replace with more relevant information (see unity/diversity

model of EF by Miyake & Shah, 1999; Miyake, Friedman, Emerson, Witzki, Howerter, &

Wager, 2000; Friedman, Miyake, Corley, Young, DeFries, & Hewitt, 2006); and (c)

Integration: bind information from multimodal sources, to achieve a given task. Thus, a core

function of the executive is to maintain and manipulate information in the episodic buffer

(Baddeley, Allen, & Hitch, 2010). As executive impairments are an established aspect of

OCD (for systematic reviews: Olley, et al., 2007; Shin, et al., 2014; Snyder, et al., 2015; Del

Casale, Rapinesi, Kotzalidis, De Rossi, Curto, Janiri, Criscuolo, Alessi, Ferri, De Giorgi,

Sani, Ferracuti, Girardi, & Brugnoli, 2016; Bragdon, Gibb, & Coles, 2018), it is expected that

they will contribute to memory impairments in two main situations:

i. In the presence of task irrelevant distractors, those with OCD are less able to inhibit

their attention to them (Coles & Heimberg, 2002), which interferes with attention-

dependent bindings in the episodic buffer, and impairs subsequent memory

performance (Gao, Wu, Qiu, He, Yang, & Shen, 2017).

ii. Those with OCD are less efficient in how they employ organizational strategies in

the presence of complex stimuli (Kuelz, et al., 2004). This will result in memory

impairments generally in visual tasks (where E and B demands are naturally high)

and for specific verbal tasks, when E and B demands are similarly high. Thus, we

expect that executive function will play a dominant role in the EBL’s moderation of

memory performance in OCD.

(2) Binding complexity (B). Binding different, multimodal features together and maintaining

these representations over time imposes a challenge that increases with the number of

features, locations and their multimodality (Fougnie & Marois, 2009). With respect to

7


binding complexity we identify two logical and empirically validated antecedents of memory

impairment in OCD: (a) the maintenance of cross-domain associations in the EB are reliant

on executively directed attention (Morey, 2009; Langerock, Vergauwe, & Barrouillet, 2014);

(b) those with OCD show gross impairments in executive functioning (Snyder, et al., 2015)

particularly at an organizational level (Kuelz, et al., 2004); (c) which results in memory

impairments for them when binding complexity is high. As such, binding complexity

removes the emphasis from the visual versus verbal distinction and places a greater emphasis

on the executive demands of the tasks, thus:

i. The inherently greater binding complexities of typical visuospatial tasks (e.g.,

multiple object-to-location bindings as observed in the RCFT) are more likely to

reveal OCD impairments than typically used verbal tasks. Complex bindings are

susceptible to interference and place greater demands upon the implementation of

correct executive control — especially when multimodal bindings are involved

(Olley, et al., 2007; Harkin & Kessler, 2009, 2011b).

ii. Verbal deficits will occur if the task relies to a similar extent upon the maintenance

of complex bindings (e.g., Wechsler Logical Memory Scale-Story Recall;

Chlebowski, 2011). This is consistent with a study by Cabrera, McNally and Savage

(2001), who reported that those with OCD relied less on organizational strategies

during the encoding of verbal information. In contrast, as simple verbal memory

tasks (e.g., word list recall) are less dependent on the maintenance of complex

bindings and are subserved by extant representations in LTM (e.g., embedded-

process model of WM; Cowan, 1999), verbal deficits are not expected to the same

extent. However, if simple verbal tasks employ OCD-threat words then this may

interfere with attention directed towards the actual task (e.g., Bohne, Keuthen,

Tuschen-Caffier, & Wilhelm, 2005; Jelinek, Rietschel, Kellner, Muhtz, & Moritz,

2012), impairing memory performance relative to neutral words.

(3) Memory load (L). If WM capacity is intact in OCD (Ciesielski, et al., 2007; Henseler, et al.,

2008; Harkin & Kessler, 2009, 2011b; Abramovitch & Cooperman, 2015), then impairments

under high load will depend on executive function (van der Wee, et al., 2003; van der Wee,

Ramsey, van Megen, Denys, Westenberg, & Kahn, 2007). An increase in load (i.e., number

of chunks to retain) places greater stress upon the correct implementation of organization

strategies (i.e., chunking), updating, and overall task-management (Smith & Jonides, 1999).

Efficient executive control reduces the overall complexity and/or load of a representation

maintained in WM (Sörqvist & Rönnberg, 2014; Simon, Tusch, Holcomb, & Daffner, 2016).

For example, when recalling a sequence of unrelated words, performance drops when the

8


number of words exceeds five or six as it is beyond the functional capacity of the

phonological loop. However, if the words create a sentence, then span can reach as high as

sixteen, which far exceeds loop capacity (Baddeley, Vallar, & Wilson, 1987). Hence,

chunking improves efficiency, as items are not individually maintained as a single unitary

representation in WM (Miller, 1956). In this understanding load has a conceptual and

empirical overlap with binding complexity. However, it differs as it refers to the increase in

cognitive load caused by more items/chunks entering WM, e.g., in sequence (e.g., n-back

tasks) and/or perhaps across space (e.g., Corsi Block-Tapping Test; Corsi, 1972) as opposed

to more multimodal features being required to be bound into a single chunk:

i. High load for visual (e.g., n-back task; van der Wee, et al., 2003) and verbal (e.g.,

WMS-Story Recall; Borges et al., 2011) tasks will similarly tax executive deficits

(e.g., chunking, updating, ordering) in those with OCD, which will then result in

memory impairments relative to controls.

ii. In contrast, as low load places less of a demand on executive function, memory

impairments will be absent (e.g., recall of words presented in sequence, Martin,

Wiggs, Altemus, Rubenstein, & Murphy, 1995) or less pronounced (e.g., digit span

task, Boldrini, Del Pace, Placidi, Keilp, Ellis, Signori, Placidi, & Cappa, 2005).

The Present Meta-Analysis – Quantification of the EBL System

As we have set out the conceptual and empirical arguments to support the EBL system, we

now outline how we standardised each dimension for the purposes of the present review. First, we

had to devise a scoring system that was not only simple but also produced meaningful ordinal

differences on each of the three EBL dimensions. For example, when a given task received a high

score for executive function (E) that it differed in obvious and pragmatic ways to a task that scored

lower on this dimension. To this end, for each dimension, we defined its primary features, identified

when general memory impairment was likely, and how it could contribute to memory impairment in

OCD. Then for each EBL dimension, we ranked each task in terms of high (3), medium (2), or low

(1) demand. This addresses a limitation observed in a previous meta-analysis in memory performance

in OCD, wherein “the classification of individual tasks was not based on reliable criteria” (Shin, et

al., 2014, p. 1127). We outline our definitional and ordinal criterions below.

Executive function. (a) Primary Features: Attention control (e.g., distractor inhibition) for

the maintenance, updating, and integration of information in WM (Wolters & Raffone, 2008). (b)

General Impairment: When task demands exceed executive function capacity and/or when a task taps

into an aspect(s) of executive function that is impaired. (c) Contribution to Memory Impairment in

OCD: When a task taps into an aspect of executive function that is impaired in OCD participants.

Examples for each of the ordinal ratings were as follows: (a) High: Tasks that require higher order

9


executive functions, e.g., deployment of organizational strategies, dual task demands. (b) Medium:

Tasks that combine a simple task with a component that distracts executive function from the primary

task, e.g., DMTS with distractor stimuli. (c) Low: Executive function serves simple maintenance of

information in WM, e.g., digit span.

Binding complexity. (a) Primary Features: Requirement to bind numerous, complex and

different (multimodal) aspects of features and maintain them in WM within space and across time

(Treisman & Zhang, 2006). (b) General Impairment: The challenge to maintain these bindings

increases with the number of features, locations and their multimodality (Fougnie & Marois, 2009).

(c) Contribution to Memory Impairment in OCD: When those with OCD fail to deploy organizational

strategies (E) to organize complex visuospatial images (and likely complex verbal information, e.g.,

WLM Story recall) into manageable parts in WM. Examples for each of the ordinal ratings were as

follows. (a) High: Complex and numerous within- and between-object location bindings and so

exceed episodic buffer capacity (i.e., approximately 3-4 feature-object-location bindings; Luck &

Vogel, 1997; Langerock, et al., 2014) and/or demand to organize and bind complex multimodal

information as manageable chunks, e.g., RCFT. (b) Medium: A number of simple object-to-location

bindings are present but are within the capacity of the episodic buffer, e.g., simple DMTS task. (c)

Low: Limited to no bindings, e.g., neutral word recall.

Memory load. (a) Primary Features: The amount of chunked information maintained in WM

at any given time (de Fockert, Rees, Frith, & Lavie, 2001). (b) General Impairment: An increase in

load (i.e., number of chunks to retain) places greater stress upon the correct implementation of

organization strategies (i.e., chunking), updating, and overall task-management (Smith & Jonides,

1999). (c) Contribution to Memory Impairment in OCD: High loads (visual or verbal) will expose the

deficits of those with OCD in efficiently reducing (i.e., via chunking, updating, ordering) the overall

complexity and/or load of a representation maintained in WM. Examples for each of the ordinal

ratings were as follows. (a) High: Complex tasks where successful performance requires efficient

chunking, updating and sorting, e.g., n-back task (Kane & Engle, 2002). (b) Medium: Moderately

complex tasks, where task performance may but is not entirely dependent on a reduction of load via

mechanisms such as chunking. (c) Low: Stimuli/task demands are such that stimuli/information can

be easily chunked/organised, and/or assisted by representations in LTM (e.g., see higher WM spans

for verbal versus spatial information; Langerock, et al., 2014), e.g., neutral word recall.

Total EBL score. It is important to note that each of the EBL dimensions does not operate in

isolation but rather are interdependent, a relationship that can sometimes be synergistic in nature. For

example, complex visual-spatial reproduction tasks have an intrinsically high binding complexity and

likely load, which places a demand on the executive functioning to use efficiently organizational

strategies/chunking to aid encoding, maintenance and recall. As such, we calculated the total EBL

10


score for each task (i.e., E+B+ L=Total EBL Score), this allowed us to examine the effect of the

overall model for EBL on memory performance between those with OCD and controls.

The EBL taxonomy creates a set of multidimensional and interrelated moderators, which

poses a challenge for traditional two-level models of meta-analysis (e.g., dependence between effect

sizes). As such the present review utilises a three-level approach (for review see Cheung, 2014),

recently used in the OCD literature by Fradkin, Strauss, Pereg, and Huppert (2018). This approach

offers us three main methodological advantages of specific relevance to the present review. First,

neuropsychological studies on OCD often use multiple measures (e.g., RCFT and California Verbal

Learning Task: CVLT; Delis, Kramer, Kaplan, & Ober, 1988) from the same participants, which

creates an issue of dependency between effect sizes. Multilevel meta-analysis avoids this issue and

allows a combination of dependent measures from different tasks from the same group of participants

(see Cheung, 2014). As noted, comparing effect sizes of tasks within the same domain with tasks

completed by the same participants provides the analysis with greater sensitivity to identify specific

deficits (Fradkin, et al., 2018). Second, Fradkin and colleagues highlighted why a meta-analysis of

the proposed EBL taxonomy is possible as they noted that cognitive tasks “often include a complex

set of scores and outcomes, and these complex structures are often difficult to integrate in quantitative

and qualitative reviews” (p. 497). This acknowledges that cognitive dysfunction in disorders such as

OCD is rarely due to one-dimensional relationships between specific cognitive components. Third,

the authors also highlighted “the importance of including different scores derived from the same task

[emphasis added] when reviewing neuropsychological and cognitive deficits [and that] multilevel

meta-analysis … allow[s] the integration of effects of complex structures” (Fradkin, et al., 2018, p.

497). This emphasises the need for- as well as the suitability of the present quantitative analysis of the

EBL taxonomy. A multi-level approach recognizes the complexity of memory impairment in OCD,

allows the use of effect sizes from a multitude of tasks from the same participants, and provides the

methodological flexibility to quantify the multi-level and interdependent set of moderators that the

EBL proposes.

Domain-specific moderator analyses and sub-task comparisons. In the OCD literature, it

is common to categorise tasks into specific domains. This allowed us to compare effect sizes for

different tasks within these domains, to offer a descriptive overview of the relationship between

domains and effect sizes. We identified eight main memory dimensions, and twenty-six subdomains

in the present review, which we now outline below (see Table 3).

Reproduction of Complex Visual Shapes. This refers to tasks, which present a complex

geometric pattern at encoding, with memory tested for the accuracy to reproduce that visuospatial

representation after a delay. Within this domain, we identified three sub-tasks, which satisfied our

criterion: Combination of Visual Reproduction Tasks (VR Tasks), RCFT, and Wechsler Memory

Scale - Visual Reproduction (WMS-VR). In Harkin and Kessler (2011a) we proposed that these tasks

11


place unique demands on each of the EBL dimensions which makes memory deficits in OCD

probable: (a) Executive Functioning: Those with OCD show consistent executive impairments (E) in

organization of complex visuospatial information (Kuelz, et al., 2004; Abramovitch, et al., 2013),

attention to details over the whole image (Cabrera, et al., 2001), and shifting from details to the whole

(Savage, Baer, Keuthen, Brown, Rauch, & Jenike, 1999). (b) Binding Complexity: successful memory

of multiple geometric shapes relies on binding, i.e., features/objects to locations, objects relative to

other objects (Hyun, Woodman, & Luck, 2009). Dysfunctional executive performance in this context

will interfere with those bindings and impair subsequent memory performance. (c) Load: high load in

these tasks is due to the executive inefficiency (E) to reduce the overall binding complexity (B) of a

complex figure by chunking it into manageable parts as opposed to isolated parts (Thalmann, Souza,

& Oberauer, 2019) that we observe in those with OCD (Savage, Deckersbach, Wilhelm, Rauch, Baer,

Reid, & Jenike, 2000) and which increases load in WM.

Span Sequence. It includes tasks that require the memorization of items in sequence and

recall of those items in that sequence. Three sub-tasks were identified: n-back (Kirchner, 1958),

symbol, and digit tasks (M.D. Lezak, Howieson, & Loring, 2004). While all these tasks rely on the

maintenance of the correct sequence of items in WM (i.e., primarily a function of executive control),

we expect to see memory impairments for those tasks which place additional pressures on executive

functioning (E) and load (L) (i.e., n-back task) compared to those that do not (e.g., digit span task).

Typically, binding complexity is low in this type of tasks.

Spatial Span. These tasks present stimuli which move across a series of spatial locations,

memory is then tested for the accurate recall of those stimuli in sequence and locations (Brown,

2016). Three sub-tasks satisfied this criterion: Tower of London – memory condition (TOL-MC;

Shallice, 1982), Self-Ordered Search Task (SOST; Petrides & Milner, 1982) and the Corsi Block-

Tapping tasks (CBTT; Corsi, 1972). Accurate performance in these tasks places a premium on the

binding, maintenance and manipulation of visuospatial representations in WM (Rojdev, Krikorian,

Feldman, Tracy, & Williams, 1998; Phillips, Wynn, Gilhooly, Della Sala, & Logie, 1999; Welsh,

Satterlee-Cartmell, & Stine, 1999).

Delayed Match-To-Sample Paradigm (DMTS; Sternberg, 1966). This paradigm presents an

encoding set and then, after a short delay, a comparison stimuli/us is presented to probe memory

(Parr, 1992). The two subtasks were: (a) Basic Storage (i.e., classic DMTS paradigm: encode, delay,

and memory probe) and (b) Storage and Distractor Inhibition (e.g., distractor/interference stimuli

presented between the encoding and memory probe; Harkin & Kessler, 2009). Consistent with a body

of literature, we expect that basic WM storage will be intact in OCD (Ciesielski, Hamalainen, Lesnik,

Geller, & Ahlfors, 2005; Ciesielski, et al., 2007; Henseler, et al., 2008). However, when a distractor is

presented, this will tap into the established executive impairments (i.e., inhibition) in those with OCD

12


(see Enright & Beech, 1993; Harkin & Kessler, 2012a) which will interfere with the stimuli-location

bindings required for accurate memory performance (see Omori, et al., 2007; Harkin & Kessler,

2009, 2011b; Harkin, et al., 2012b).

Recall of Simple Verbal Information. These tasks present verbal information (i.e., words) and

then require participants to recall it. We identified three sub-tasks, i.e., Threat and Neutral Word, and

Cued recall (i.e., the first two letters of a word provided as a cue). We expect that threat-related words

will interfere (i.e., via an executive dysfunction in the ability to inhibit irrelevant/threat stimuli;

Clayton, Richards, & Edwards, 1999; Cohen, Lachenmeyer, & Springer, 2003) with the maintenance

of words for the main memory task.

Recall of Complex Verbal Information. The sub-tasks were a Combination of Complex

Verbal Tasks, Wechsler Memory Scale – Logical Memory – Story Recall (WMS-Story Recall),

CVLT, Rey Auditory Verbal Learning Task (RAVLT), and Auditory Verbal Learning Tasks

(AVLT). As in the previous domain, these tasks present verbal information to participants that they

have to recall at a later point. However, successful performance on these tasks required participants

navigate an additional layer of complexity. For example, in the combination category, we included an

effect from Grisham and Williams (2013) who employed the Operation-Span Task (OSPAN; Turner

& Engle, 1989). This task required the memorisation of a set of unrelated words while performing a

series of math equations, memory performance was the identification of the correct words in the

correct location. Thus, in comparison to tasks in the previous simple verbal domain, the overall EBL

demand is greater. In that: (i) executive control is operating under dual-task conditions (i.e., primary

word task and secondary math equations), (ii) binding complexity is present (i.e., words maintained

in locations), and (iii) load is greater (i.e., due to the preceding E and B demands). Similar to the

RCFT, we therefore expect that complex verbal tasks will score high on overall EBL and E demands.

Recognition Memory. This included tasks that present memory items (i.e., originally

presented as the memory task) along with new items. Accurate performance requires the participants

to correctly identify the memory items while ignoring/rejecting new/distractor items (Moreno-

Castilla, Guzman-Ramos, & Bermudez-Rattoni, 2018). Three sub-tasks satisfied this definition, i.e.,

Visuospatial, Object and Verbal recognition. We expect a classic pattern of impairment in OCD, with

greater impairment for visuospatial than verbal recognition memory, a difference we suggest can be

explained by greater EBL generally and E contribution specifically for the former versus the later

tasks.

Declarative and Implicit/Procedural Memory. For this, we included Prospective (i.e.,

remember to perform an action; McDaniel & Einstein, 2007), Source (i.e., recall source of learned

information; Pandey, 2011) and False (i.e., recall information that has not been studied; e.g., Dees-

Roedigier-McDermott paradigm; Roediger & McDermott, 1995), and Procedural/Motor memory

(i.e., learning without awareness on tasks dependent on motor performance; e.g., star maze learning:

13


see Mataix-Cols, 2003). Generally, procedural/motor memory exhibits limited to no impairment in

OCD (see Mataix-Cols, 2003; Roth, Baribeau, Milovan, O'Connor, & Todorov, 2004; Shahar,

Teodorescu, Anholt, Karmon-Presser, & Meiran, 2017). As these domains are novel to meta-analysis,

they provide an exploratory overview of their EBL demands and relationship to effect sizes.

In sum, we argue that this review aims to answer a point raised by Greisberg and McKay

(2003): “that a model of neuropsychological functioning in OCD must be articulated if progress is to

be made in delineating specific deficit areas” (p. 112). In addition to this, we also examined a range

of methodological, clinical, demographic and sample characteristics as potential moderators of effect

sizes.

Methods

Protocol

The Cochrane Collaboration’s (Moher, Liberati, Tetzlaff, & Altman, 2009) general

guidelines for conducting systematic reviews and meta-analyses were followed during each stage of

the review process. The review has been reported in accordance with the PRISMA (Moher, et al.,

2009) guidelines, and the flow-chart can be found in the supplementary materials.

Inclusion and Exclusion Criteria

Included studies were required to compare memory performance on at least one task between

adults with OCD or OCD-type traits (e.g., checking) and healthy controls. Studies examining visual,

verbal, and WM were all included. Participants with subclinical OCD were included and coded

accordingly. Participants with acquired OCD (e.g., following head injury) were excluded, as this

group is generally considered clinically distinct from idiopathic OCD groups and illness will in most

cases be associated with neurological injury (Coetzer, 2004). As hoarding disorder has recently been

associated with different neuropsychological deficits than OCD (Tolin, Villavicencio, Umbach, &

Kurtz, 2011), patients with this as a primary diagnosis were also excluded from the analysis. Healthy

controls were defined as adults without any reported neurological deficits, and who were not

exclusively diagnosed with another mental illness (e.g., depression), and who were not related to the

OCD participants (in order to avoid tapping into a potential OCD endophenotype). Correlational

studies were included if it was possible to obtain data to facilitate effect size calculation. Treatment

studies were included if there were baseline memory scores available. Studies reporting only the

Wisconsin Card Sorting Test were excluded, as this task is thought to mostly tap into cognitive

functions other than memory (e.g., set-shifting), making it too complex of a test to directly access

memory performance alone (Stratta, Daneluzzo, Prosperini, Bustini, Mattei, & Rossi, 1997). Studies

had to be available in English.

14


Search Strategies

Searches were conducted between October 2018 until March 2019. The search terms used to

access literature was ‘(wash* OR check* OR hoard* OR obsessive-compulsive* OR OCD OR

clean*) AND (executive OR bind* OR load* OR visual OR verbal) AND (memory)’. Keywords were

developed based on previous literature e.g., Leopold and Backenstrass (2015), and agreed upon by

the first and second author. Studies were searched via the following social scientific databases: Web

of Science, Psychinfo, Medline, PubMed, OVID, CINAHL, PsychArticles. ProQuest Theses was

searched to identify potential grey literature, and prominent authors (most prolific 10% as determined

by the initial literature search) in the field were also contacted enquiring about unpublished material.

All authors contacted about additional data (see information below for further details), were asked

about any unpublished material they might have. Where information to facilitate effect size

calculation was missing, authors were contacted requesting additional data. To this end, we contacted

10 authors, five of which provided the requested data. Search hits were converted to RIS-files (Clark,

2019, personal communication), and imported into the web application Rayyan (Ouzzani, Hammady,

Fedorowicz, & Elmagarmid, 2016), where they were screened for eligibility. Rayyan is a freely

available review tool which facilitates the screening of titles and abstracts, and allows the researcher

to attach inclusion decisions for each of the entries, whilst also customising the label behind the

decision (e.g., “No OCD group”; “Paediatric population” etc.,). A total of 7267 studies were

identified through database searching. Full texts of 321 articles were examined, resulting in 144 to be

included in the final review. The Flow-chart can be found in the supplementary material.

Data Extraction and Coding

A data extraction sheet was developed by the first and second author. The data extracted were

as follows: (a) basic study information: name, year, and authors of publication, publication status,

country of origin (1=USA/Canada, 2=UK/Europe, 3=Asia, 4=other); (b) study and effect size

identification: study ID, effect size ID; (b) effect size measures: control N, OCD group N, effect size,

standard error, variance; (c) demographic participant info: percentage of females, mean age; (d)

clinical characteristics: whether YBOCS was administered, YBOCS score, mean age of OCD onset,

whether or not sub-clinical OCD participants were included, OCD severity (1=sub-clinical; 2=low;

3=moderate; 4=severe), OCD sub-type (1=washers, 2=obsessions, 3=checkers, 4=others/undefined;

comorbidities in OCD group (1=comorbidities present in all, 2=no comorbidities recorded, 3=mixed,

4=not measured), did the OCD group receive therapy at the time of the study (1=yes all participants,

2=no participants, 3=mixed treatment, 4=not established); (e) task characteristics: whether task was

visual or verbal, combined and individual EBL scores; and (f) published in peer reviewed journal

(1=yes, 2=no); and (g) methodological quality. The methodological quality score was adapted from

Leopold and Backenstrass (2015), and included items on ethical approval, sample description, info on

15


OCD severity and co-morbidities, reliability of OCD diagnosis and neuropsychological tests, number

and training of test supervisors, details on OCD treatment, and control group matching. Scores on

each item ranged from 1 (no information) to 3 (comprehensive information).

To obtain EBL scores, all tasks were coded on three individual components; executive load,

binding complexity, task load. A detailed explanation of the theoretical background to each of these

components can be found in previous sections of this paper, and in Harkin and Kessler (2011a). Tasks

were scored 1 through to 3, depending on how much demand was placed on the individual

component: 1 indicated low to little demand, 2 moderate demand, and 3 indicated high demand. EBL

scores could therefore vary between 4 and 9. Initially tasks were also assessed on emotional valence

and ecological validity, but as they did not generally vary across tasks (16 and 10 effect sizes,

respectively, scored above 1), they were excluded. A key consideration of the current examination is

therefore how memory performance between patients with OCD and healthy controls (as measured

by the Cohen’s d) vary according to task demands. For reliability purposes, 10% of the data were

coded blindly and independently by a second coder. Agreement between coders was high at 98%.

Any disagreements were resolved through discussion and final agreement within the team, with

virtually all discretions attributable to slight variation in calculations used.

Statistical Analysis

Cohen’s d was calculated as measure of effect size of the difference in performance on each

of the memory tasks between OCD participants and healthy controls. A positive effect size indicated

a memory deficit among participants with OCD, compared to control participants. Where Cohen’s d

was not reported in the original study, effect sizes were either converted from other effects (e.g., F) or

calculated manually using the Campbell Collaboration’s Practical Meta-Analysis Effect Size

Calculator (Wilson, 2019). As recommended by Assink and Wibbelink (2016), the variance was

calculated as SE^2.

A number of studies provided more than one effect size, as participants had been

administered several memory performance tasks during the study period, thus violating the normal

requirement of independent effect size measures in meta-analysis (Rosenthal, 1986; Cheung, 2014).

Dependency of effect sizes normally means that effect sizes within studies are correlated (as these are

expected to show a certain similarity); this creates an overlap of information and inflates information

produced by the analysis, which can result in an over-confidence in its results (Van den Noortgate,

López-López, Marín-Martínez, & Sánchez-Meca, 2013; Assink & Wibbelink, 2016). Whilst it is

possible to conduct sub-group analysis or aggregating effect sizes , this reduces the number of effect

sizes analysed in a set, therefore limiting power of the analysis, something that is a particular concern

when conducting multiple moderation analyses (Assink & Wibbelink, 2016). In the present review,

16


these solutions would not have been suitable, since a main aim was to examine the moderating effect

of task characteristics, across different memory domains/tasks.

Where correlations between effect sizes are not known, it is possible to fit a three-level meta-

analytical structure. This analysis considers three levels of variance components distributed across the

model, including variance between effect sizes from the same study and variance between studies;

this therefore allows for an examination of how effect sizes vary between participants (level 1),

outcomes (level 2), and studies (level 3) (Cheung, 2014). This type of approach produces a robust

analysis, and has been successfully implemented in recent meta-analytical research into OCD and

cognition (Fradkin, et al., 2018) and was the approach followed here.

The current analysis was conducted using the rma.va function in the Metafor package for the

statistical software environment R (R Core Team, 2013); R Core Team, 2014), and recommendations

of Viechtbauer (2010). A mixed-effects model was fitted, and estimation was based on the restricted

maximum likelihood estimator. The analysis examined the variance distribution over the three levels,

the overall effect (i.e., memory performance of those with OCD compared to controls), and the

effects of a number of moderating variables. As recommended by Hox (2010) and Assink and

Wibbelink (2016), moderators were first examined individually, and then combined into one analysis.

This allows for initial significance screening, whilst also accounting for the possibility of variables of

interest being intercorrelated, producing multicollinearity in analyses. Moderators are reported with

mean effect sizes, significance levels, and confidence intervals. R Code was adapted from Assink and

Wibbelink (2016) and Harrer, Cuijpers and Ebert (2019). Visuals were created using ggplot2

(Wickham, 2016). Assessment of methodological quality occurred by entering it as a moderator of

effect sizes. Publication bias was assessed using the funnel plot function (funnel) in R, as

recommended by Harrer et al. (2019), and Egger’s regression coefficients (Egger, Smith, Schneider,

& Minder, 1997).

Results

Data Preparation

Based on recommendations by Snyder et al. (2015), effect sizes 3 SD above or below the

mean effect size (d=0.50) were considered outliers, and thus excluded. On the basis of this, four

effect sizes relating to two studies were removed: one relating to RCFT (Boldrini, et al., 2005), and

three relating to word learning (Irak & Flament, 2009)). As recommended by Assink and Wibbelink

(2016), categorical moderators were dummy coded (0: absent; 1: present), to allow for an estimation

of mean effects of each category. The categorical moderator was whether the task was visual or

verbal.

Preliminary Analyses

17


In total, 144 independent studies were included, totalling 305 effect sizes. Thus, the mean

number of effect sizes for each study was 2.13. Those included for meta-analysis are indicated with a

* in the reference section. These studies originated from a variety of countries: USA (N=33),

Germany (N=19), Korea (N=17), India (N=11), Spain (N=10), Turkey (N=9), Italy (N=6), Australia

(N=6), France (N=6), Canada (N=5), China (N=4), Japan (N=3), Austria (N=2), Netherlands (N=2),

Iran (N=2), Sweden (N=1), Switzerland (N=1), Denmark (N=1), South Africa (N=1), Brazil (N=1).

Across all studies, these included 9200 participants (min=18, max=355; median=60). Of these, there

were 4424 OCD patients (mean age=31.68; 52.65% women) and 4776 healthy controls (mean

age=31.06; 52.71% women). Two t-tests confirmed that there was no significant differences in age

and sex between the two groups. The vast majority (N=133) of studies included patients who had

been formally diagnosed with OCD, and 13 included those with sub-clinical OCD. Patients were

mostly (N=119) diagnosed with the YBOCS. The remaining studies (N=27) utilised a variety of other

measurements. Seventy-four studies included samples where all, or some of the OCD patients were

medicated, whereas 33 studies included un-medicated patients only. Thirty-seven studies did not

report the medications status of the participants. There was an even split between visual (N=143) and

verbal (N=148) tasks, with a small minority (N=15) combining visual and verbal elements. The mean

EBL score across the individual tasks was 6.89 (median=7, min=4, max=9). This indicates that

overall, tasks across the sample placed considerable demand on executive function, binding

complexity, and memory load.

Main and Heterogeneity Analyses

The first step of the analysis estimates the overall effect size for the memory difference

between those with and without OCD, including 305 effect sizes from 144 individual studies. Across

all studies the overall mean effect for the memory deficit of patients with OCD as compared to

healthy control was medium-sized, d=0.50, SE=0.03, 95% CI=[.43, .57], p<.001). The second step of

the analysis estimated the difference between within- (level 2) and between-study (level 3) variance

components, an important aspect of a three-level meta-analysis. This is assessed through two separate

log-likelihood-ratio tests, where the original model (with freely estimated variance at level 2 and 3

respectively) is compared to one where the variance at each of the level is fixed. The analyses

suggested that there was significant variability (ps<.001) between effect sizes (level 2), and also

between studies (level 3), indicating that moderator analyses should be conducted (Assink &

Wibbelink, 2016). Based on formulas by Cheung (2004), the total variance distribution is as follows:

level 1: 28.72%; level 2: 31.88%; level 3: 39.39%. As recommended by Assink and Wibbelink

(2016), moderation analyses should be conducted if less than 75% of the variance can be attributed to

level 1.

18


EBL moderator analyses. A number of task-related characteristics were tested as

moderators of OCD memory performance, including the EBL framework and whether the task was

visual or verbal in nature. Initially, individual moderator analyses were conducted on task

characteristics, which included how the task was characterised on the EBL-framework (combined

EBL score, and an individual assessment of executive function, binding complexity, and load,

respectively), as well as whether the task was visual or verbal in nature. The overall model for EBL

was significant: F(1, 303) = 38.07, p<.001, indicating that how a task is classified on the EBL-

framework moderates the overall difference in memory performance between those with OCD and

healthy controls. Specifically, as EBL demand increases, those with OCD performed worse on

memory tasks than healthy controls: β=0.11, p<.001, 95% CI=[0.08, 0.14]. Table 1 provides the

outputs of the main moderation analyses.

Table 1.Main Moderator Analyses

Variable k d (se) p C-, C+ Q (p)Main Analysis 305 0.50 (0.03) <.001 0.43, 0.57 943.86(<.001)

EBL ModelFull EBL model 305 0.11(0.02) <0.001 0.08, 0.15 884.79(<.001)Executive function 305 0.22(0.30) <.001 0.16, 0.28 845.90(<.001)Binding Complexity 305 0.15(0.03) <.001 0.09, 0.21 913.64(<.001)Memory Load 305 -0.13(0.05) <.001 -0.23, -0.03 933.10 (<.001)

Type of TaskVisual/Verbal 290 0.20 (0.05) <.001 0.10, 0.30 868.07 (<.001)

Participant CharacteristicsYBOCS 277 0.01(0.01) .09 -0.002, 0.03 873.78 (<.001)OCD group % women

275 <.001(<.001) .96 -.003, 003 784.16 (<.001)

OCD group age 288 0.02(<.001) .002 0.01, 0.3 809.12 (<.001)

Individual EBL moderator analyses. As for the individual components of the EBL

framework, executive function was a significant moderator on memory performance, β=0.22, p<.001,

95% CI=[0.16, 0.27]. Specifically, as executive function demand increased, so did the memory deficit

for OCD participants. Additionally, increased binding complexity of the memory task also increased

OCD memory deficit: β=0.15, p<.001, 95% C =[0.09, 0.21] Interestingly, the moderating effect of

memory load went in the opposite direction, with increases in load leading to better memory

performance of those with OCD as compared to healthy controls, β=- 0.13, p<.001, 95% CI=[-0.23, -

0.03]. Figure 2 provides visual plots of the moderating effect between the EBL model and individual

components (i.e., executive function, binding complexity, memory load) and memory performance in

OCD.

19


Figure 2. Visualisation of Individual Moderation Effects.

Visual versus verbal moderator analysis. A moderator analysis was then performed on the

type of memory task, which was classified as either visual or verbal in nature. Due to their relative

scarcity, memory tasks that combined visual and verbal elements were excluded, leaving a final

sample of 290 effect sizes. Verbal tasks were classified as the reference category, and visual tasks

tested against this. The overall effect sizes for visual and verbal tasks was d = 0.65 and 0.44,

respectively. The full model was significant; F(1, 288)=14.62, p<.001, with a greater memory deficit

among those with OCD when visual tasks were used, with a mean effect of 0.20, se=0.05, 95%

CI=[0.10, 0.30]. However, Figure 3 (left panel) illustrates the moderation analysis between type of

task and the EBL model. It suggests that visual tasks place a greater demand on the combined aspects

of the EBL, which appears to result in a greater memory difference between those with OCD, and

those without. It is possible this goes some way towards explaining why generally, those with OCD

perform worse than controls on visual tasks, but not always on verbal tasks.

20


Combined moderator analysis. Variables relating to task characteristics were then

combined into one analysis, as per the recommendations by Hox (2010). As it is expected that there

will be a degree of confound among the task characteristics, this allows for an examination of

whether individual moderators remain significant when examined together. To avoid running an

analysis with an excess of variables, the EBL framework and whether the task was visual or verbal

was combined with only the strongest individual EBL aspect, which was executive function. The full

model was significant, F(1,286)=19.33, p<.001. Importantly, only executive function remained a

significant moderator in this context, β=0.23, 95% CI=[.10, 0.36]. Therefore, the effects of whether

the task is visual or verbal, as well as the full EBL-framework were no longer significant moderators

when tested in the context of executive function, which appears to hold the main task-related

moderating impact on the memory deficit between those with OCD, and healthy controls. It therefore

appears that executive function is the driving mechanism behind the EBL framework’s impact on

OCD memory performance. As is plotted in Figure 3 (right panel), visual tasks appear to place a

Figure 3. Left Panel: Moderating Effect of EBL Model and Type of Task on the OCD Memory Deficit. Right Panel: Moderating Effect of Executive Function and Type of Task on the OCD Memory Deficit. considerable demand on executive function, which results in more pronounced memory differences

for those with OCD.

Domain-specific moderator analyses. Table 2 provides an overview of the individual

moderation analyses for the main memory domains. We only entered executive function into the

analysis, as in the previous full moderation model it was the only dimension to remain significant.

Table 2

Executive Function Moderator Analyses per Memory Domain Memory Domain k B p C-, C+ Q(p)Reproduction – Complex Geometric 56 0.78 <.001 0.66, 0.90 135.32(<.001)Span –Sequence 38 0.62 <.001 0.36, 0.87 115.81 (<.001)

21


Span – Spatial 26 0.13 .21 -.08, .03 37.78 (.04)Delayed Match-To-Sample 16 0.20 .10 -0.04, 0.43 14.00(.45)Recall – Simple Verbal 19 -0.28 .09 -0.61, 0.05 13.80(.68)Recall – Complex Verbal 53 0.30 .02 0.05, 0.53 164.72 (<.001)Recognition 21 0.16 .22 -0.10, 0.41 26.97 (<.02)Declarative Memory 12 .008 .45 -0.15, 0.30 15.93(<.10)Implicit Memory 4 0.69 .21 -.93, 2.31 3.64 (.21)

In line with the overall analysis, executive function remains a considerable moderator in tasks

that loads more heavily onto this dimension, e.g., complex visual reproduction tasks. Whereas, tasks

that place little demand on executive function (e.g., implicit and declarative memory tasks), it does

not moderate effects, importantly, they also generally report less differential memory performance

between those with OCD, and those without.

Task domain comparisons. For illustration purposes, Table 3 (see Supplementary materials

for included studies) provides EBL scores and effect sizes for individual tasks within each memory

domain. Broadly it illustrates that memory domains with a higher EBL score (and specifically

executive function) report higher effect sizes for the memory differences in those with OCD, as

compared to healthy controls. With respect to the memory performance of those with OCD to

controls, we highlight the following. First, in accord with the parameters of the EBL taxonomy the

reproduction of complex geometric shapes produces large to medium impairments across the three

main measures of combined VR tasks (d=0.88), RCFT (d=0.86), and WMS-VR (d=0.54). Second, the

specific type of sequence recall appears to make a difference in the memory performance of those

with OCD, with very large (d=1.13), medium (d=0.67) and small (d=0.31) effect sizes observed for

the n-back, symbol, and digit tasks, respectively. For spatial span tasks, similar medium effect sizes

(d = 0.74, 0.64, 0.50) were observed across the three subtasks (i.e., TOL, SOST, CBTT). Third, for

the pure WM domain, basic storage shows a very small effect size (d=0.14) compared to when basic

storage was combined with a distractor task/stimulus which resulted in a medium effect size (d=0.62).

This suggests that WM capacity is generally intact but impairment in OCD will occur when a

distractor task/stimulus is simultaneously present. Fourth, the recall of threatening words was

associated with a medium effect size (d=0.46) compared to small to negligible effect sizes for cued

(d=0.22) and neutral (d=0.02) word recall. Fifth, the recall of more complex verbal information

shows larger effect sizes (d=0.97) compared to a medium effect for story recall (d=0.65). Medium to

small effect sizes (d=0.37, 0.36, 0.17) were then observed for the similar tasks of the CVLT, RAVLT,

and AVLT. Sixth, the recognition of visuo-spatial information had a medium effect size (d=0.57) in

contrast to the small effect sizes for object (d=0.27) and verbal (d=0.20) recognition. Lastly, across

declarative and implicit memory, we saw a range of medium to no effect sizes for prospective,

source, false and procedural/motor memory, i.e., d=0.43, 0.30, 0.10, -0.04, respectively.

Table 3

22


Average Effect Sizes across Memory Domains and Sub-TasksMemory Domain Tasks k EBL

TotalE B L d+

Reproduction – Complex Geometric

VR Tasks1 6 8.00 3.00 3.00 2.00 0.88RCFT2 40 8.00 3.00 3.00 2.00 0.86WMS-VR3 10 8.00 3.00 3.00 2.00 0.54

Span – Sequence N-Back4 5 7.40 2.40 2.00 3.00 1.13Symbol5 5 5.80 1.40 1.40 3.00 0.67Digit6 28 5.04 1.04 1.04 2.96 0.31

Span – Spatial TOL7 3 8.00 3.00 2.00 3.00 0.74SOST8 10 8.78 3.00 2.70 3.00 0.64CBTTs9 13 6.69 2.23 1.46 3.00 0.50

Delayed Match-To-Sample

Storage + Distractor Inhibition10a

7 7.43 2.71 2.29 2.43 0.62

Basic Storage10b 9 6.67 1.67 2.33 2.67 0.14Recall – Simple Verbal

Threat Words11 3 6.67 2.00 1.67 3.00 0.46Cued12 4 6.00 1.75 2.00 2.25 0.22Neutral Words13 12 5.83 1.83 1.83 2.17 0.02

Recall – Complex Verbal

Complex Verbal Tasks14

10 6.80 2.40 1.40 3.00 0.97

WLM-Story Recall15 13 7.83 3.00 2.00 2.83 0.65CVLT16 15 7.60 2.00 2.87 2.73 0.37RAVLT17 11 7.00 2.00 2.00 3.00 0.36AVLT18 9 7.00 2.00 2.00 3.00 0.17

Recognition Visuo-Spatial19 8 6.88 2.38 2.00 2.50 0.57Objects20 4 6.25 2.00 1.50 2.75 0.27Verbal21 9 5.89 1.67 1.56 2.67 0.20

Declarative Memory

Prospective22 5 6.40 2.40 1.80 2.20 0.43Source23 3 7.00 3.00 2.00 2.00 0.30False24 4 6.25 2.00 1.75 2.50 0.10

Implicit Memory Procedural/Motor25 4 7.25 2.75 2.00 2.50 -0.04

Participant characteristics. Moderation analyses for the participant characteristics

moderation model (percentage of women, participant YBOCS score, age) were non-significant, apart

from the analysis on age. Specifically, as age increased, those with OCD performed worse on

memory tasks than healthy controls: β=0.02, p<.01, 95% CI=[0.01, 0.03], although it should be noted

that the effect was very small.

Study Quality and Publication Bias

All studies were given a methodological quality score, with a mean quality score for the

overall sample of 21.94 (median=23.00, min=2, max=29). The potential maximum score was 30,

indicating an overall good quality of methodology of included studies. To examine whether study

quality was associated with overall result, a moderator analysis with the methodological quality score

indicated that the methodological quality was not associated with overall results: β=0.01, p=0.24,

95% CI=[-0.007, 0.03]. Although there is considerable symmetry among the majority of studies, a

small number fell outside of the funnel (see supplementary materials). The Egger’s regression

coefficient was significant: β= .60, p<.01, 95% CI=[0.52, 2.68]. In this instance, Duval and

Tweedie’s (2000) trim and fill approach is suggested, however, it has been observed to drastically

23


underestimate effect sizes when between-study heterogeneity is large (Peters, Sutton, Jones, Abrams,

& Rushton, 2007). Heterogeneity in the present meta-analysis we attribute to variance in sample sizes

across studies (i.e., total n: 20 to 410 participants; Simonsohn, 2017), a large number of studies

(k=144) and numerous interrelated effect sizes (i.e., 305 data points) for 28 individual memory tasks

from 9 memory domains. In addition, the within-study sampling variance of 28.72% fell well below

the acceptable threshold of 75% to conduct subsequent moderator analyses on effect sizes as

originally extracted (Assink & Wibbelink, 2016). Thus, we did not conduct additional

transformations on our data, especially when our original effect sizes (Table 3) were comparable to

those reported in previous meta-analysis of memory performance in OCD (see Abramovitch, et al.,

2013; Shin, et al., 2014; Leopold & Backenstrass, 2015; Snyder, et al., 2015).

Discussion

Memory performance in OCD is a commonly researched aspect of this disorder, which has

made it the subject of numerous selective (Greisberg & McKay, 2003; Kuelz, et al., 2004; Muller &

Roberts, 2005; Olley, et al., 2007; Abramovitch & Cooperman, 2015) and meta-analytic reviews

(Abramovitch, et al., 2013; Shin, et al., 2014; Leopold & Backenstrass, 2015; Snyder, et al., 2015).

Despite these excellent reviews and associated research, the field has yet to provide a unified and

coherent model to understand memory impairment in OCD. This, in part, is attributable to an

emphasis on memory performance between specific memory domains (e.g., visual versus verbal) and

associated tasks (e.g., CVLT versus RCFT, respectively). As a solution, the present meta-analysis

takes a novel approach to memory impairment in OCD, wherein we standardise specific task features

as set out in the original EBL classification system (Harkin & Kessler, 2011a). We observed that the

EBL taxonomy had explanatory power for several aspects of memory performance and deficits in

OCD.

Predictive validity of the EBL model. The EBL model had predictive validity for memory

performance in OCD. Specifically, as EBL demand increases, those with OCD had poorer memory

performance relative to controls. We observed a medium-sized (d=0.50) memory deficit in those with

OCD, which is comparable to the overall effect sizes reported in previous meta-analyses on memory

performance in OCD (Abramovitch, et al., 2013; Shin, et al., 2014; Snyder, et al., 2015). The total

number of data points (e.g., 305 effect sizes) which contributed to our overall EBL model moderation

analyses adds to our confidence in our results. Together, these points support the assertion that the

pattern of memory impairments we observe are not due to spurious coding and/or study inclusion but

rather the relationship between specific dimensions of the EBL system and memory performance in

those with OCD. This gives us confidence that the present results have gone some of the way to

24


satisfy a point raised by Greisberg and McKay (2003), in that we have a model of memory

performance in OCD, that allows us to delineate deficits in specific areas.

Individual EBL components. One of the strengths of the present approach is that we were

able to isolate the effects of each EBL dimension on memory performance. Our results indicate that

as executive function and binding complexity increases, so does memory impairment. In explanation,

binding of complex stimuli/information relies upon attention to the encoding, maintenance, and

retrieval of object-object and object-location bindings (Hinton, et al., 1986; Morey, 2009; Langerock,

et al., 2014). However, when this attention is interfered with and/or insufficient to the demands of the

task, then memory impairment follows (Fougnie & Marois, 2009). In line with a body of research

(e.g., Head, Bolton, & Hymas, 1989; Enright & Beech, 1993; Bohne, Savage, Deckersbach, Keuthen,

Jenike, Tuschen-Caffier, & Wilhelm, 2005; Penades, Catalan, Andres, Salamero, & Gasto, 2005) this

identifies executive impairments as an established feature of OCD (see Snyder, et al., 2015 for a

meta-analysis on gross EF impairments in OCD), that memory impairments are secondary to

executive dysfunction (e.g., Olley, et al., 2007), and stimuli that are high in binding complexity will

expose the deficits of those with OCD to encode such stimuli/information (visual or verbal) in an

efficient manner (VanRullen, 2009).

In contrast, our findings with respect to memory load were contrary to our expectations, i.e.,

increases in load lead to better memory performance in the OCD group. To explain this, we highlight

a critique of cognitive research in OCD by Ouimet, Ashbaugh and Radomsky (2019): “methods are

rarely process-pure and often conflate cognitive processes with measure outcomes … as if the

outcome and the underlying cognitive process are one and the same” (p. 24). Therefore, our findings

may reflect a discrepancy between how we conceptually defined load (i.e., a function of stimulus

complexity due to E and B demands; Simon, et al., 2016) and then actually scored a task with respect

to load, e.g., more as a function of basic and isolated WM capacity (de Fockert, et al., 2001). This

unexpected finding might also highlight existing methodological issues with certain tasks, where

increasing the number of simple unimodal items within normal capacity limits might result in contra-

intuitive results, as it does not impose any strain on memory operation in OCD samples. In any case

this unexpected finding highlights the benefit of quantifying the impact of individual task dimensions

on memory performance, as collapsing across such dimensions will obscure their unique impact on

memory performance.

Visual versus verbal tasks. Consistent with a body of literature (Abramovitch, et al., 2013;

Shin, et al., 2014; Snyder, et al., 2015), we observed that OCD participants had greater impairment in

visual (d=0.65) compared to verbal tasks (d=0.44), and had the classic effect size difference for the

RCFT (d=0.86) and the CVLT (d=0.37) (Abramovitch, et al., 2013). However, we conclude that

visual tasks place greater demands on OCD patients than verbal tasks in concordance with the

combined EBL model , which goes some of the way to explain why those with OCD generally

25


perform worse than controls on visual tasks, but not always on verbal tasks (see reviews by Muller &

Roberts, 2005; Olley, et al., 2007; Abramovitch & Cooperman, 2015). We further conclude that

executive function likely explains this pattern. In that, the performance of verbal tasks (e.g., word

lists in CVLT) and the maintenance of verbal information in WM are likely supported via existing

representations in LTM (see Embedded Process Model of WM; Cowan, 1999), which tempers the

incorrect deployment of executive functions/attention to task demands that we commonly observe in

OCD (for review see Collette, Van der Linden, & Ponceret, 2000). In contrast, such bottom-up LTM

representations are not so readily available to support the maintenance of novel visuospatial

representations in WM (e.g., geometric shapes in the RCFT). Rather, the veridicality of fragile object-

location bindings in WM are dependent on focused and uninterrupted executive functions/attention

(Allen, et al., 2006; Fougnie & Marois, 2009). We underline the significance of executive function

demand across visual and verbal memory performance in OCD in the next section.

The importance of Executive Function. Executive function was one of the strongest and

theoretically most interesting predictors of memory performance in those with OCD. First, a pivotal

finding was that executive function negated the impact of the visual versus verbal memory task

difference in those with OCD. This later finding is important as it highlights that it is the executive

demands of a task and not the visual or verbal description of a task, which determines memory

performance in OCD. This validates an insight of Leopold and Backenstrass (2015) who drew a

relationship between an impairment in applying executive strategies to efficiently encode visual and

verbal information and subsequent memory deficits (see Savage, et al., 1999; Deckersbach, Otto,

Savage, Baer, & Jenike, 2000; Shin, et al., 2014). Second, executive function was the strongest

predictor of memory performance beyond the combined EBL model, and individual binding and load

dimensions, indicating that executive function is the driving mechanism behind the EBL’s impact on

memory performance in OCD. This was expected based on the original EBL conceptualisation by

Harkin and Kessler (2011a), where “E” was the dominant dimension in the interdependent EBL

model. Nevertheless, conforming to the oblique dimensioning of EBL, binding complexity as a

reflection of the multimodality of memory chunks, might additionally contribute towards explaining

OCD memory deficits in certain tasks (Table 2), especially those where complex multimodal

representations increase demands on executive function. Importantly, our current finding serves to

explicitly quantify the often cited observation that memory impairment in OCD is in fact secondary to

executive dysfunction (Greisberg & McKay, 2003; van der Wee, et al., 2003; Kuelz, et al., 2004;

Olley, et al., 2007; Abramovitch, et al., 2013), and highlights the importance of our novel coding and

multi-level approach. In that, if we had focused on the traditional visual-verbal distinction, our

analysis would not have uncovered the subtle, underlying and significant impact of executive

function across a range of tasks generally and for the visual-verbal distinction specifically.

26


Memory Domains. We conducted domain-specific moderator analysis with executive

function only. In general, for the domains that load heavily on executive function (e.g. reproduction

of complex visual stimuli, span-sequence, and recall of complex verbal information), the memory

performance of those with OCD was impaired relative to controls. In contrast, for domains (e.g.,

implicit and declarative memory domains) that loaded less heavily on executive function, then there

was less of a pronounced difference in memory performance between OCD participants and controls.

It is interesting to note that the executive demands of a given domain moderated effect sizes more

distinctly than a general dissociation between the visual and verbal domain. For example, executive

function moderated poorer memory performance in OCD patients on the reproduction of complex

visual images and recall of complex verbal information but not on the DMTS, spatial-span or recall of

simple verbal information domains.

Within each of these main memory domains, we categorised relevant tasks and averaged

effect sizes accordingly (Table 3). This provided a further descriptive level of analysis to that of the

previous domain-specific moderator analyses. First, the effect sizes for the visual-reproduction,

RCFT and the WMS-VR tasks was comparable to other meta-analytic reviews (e.g., Abramovitch, et

al., 2013; Shin, et al., 2014; Leopold & Backenstrass, 2015; Snyder, et al., 2015). Considering the

strong moderating effect of executive function for these tasks, we conclude that OCD patients fail to

organize complex geometric shapes in an efficient manner during encoding (see Penades, et al.,

2005). Second, for the tasks that made up the span-sequence domain, we observed a range of effect

sizes for the n-back (d=1.13), symbol (d=0.67) and digit (d=0.31) tasks. We highlight the reliability

of these findings, as a meta-analysis by Snyder et al. (2015) also reported the largest impairment of

those with OCD on the n-back, and the small effect size for digit span matches that of a meta-analysis

by Shin et al. (2014). In the domain of spatial span, we observed medium sized memory deficits in

OCD participants in the TOL (d=0.74), SOST (d=0.64) and CBTT (d=0.50). This pattern and

magnitude of effect sizes for the TOL and CBTT were similar to those reported by Shin et al. (2014).

Based on our inclusion criteria for these tasks, we conclude that those with OCD suffer from a

general impairment on the maintenance and potential manipulation of visuospatial representations in

WM.

We draw support for the argument that WM capacity is intact in OCD (Ciesielski, et al.,

2005; Ciesielski, et al., 2007; Henseler, et al., 2008; Exner, Kohl, Zaudig, Langs, Lincoln, & Rief,

2009; Abramovitch, et al., 2013) from the fourth domain that included two similar DMTS tasks.

Specifically, we observed that for a basic storage task there was little difference between OCD

patients and controls (d=0.14). In contrast, when the same DMTS task has a distractor stimulus

between encoding and the memory probe, then those with OCD suffered from considerable memory

impairment (d=0.62). This helps to underscore the point that any significant deficits in memory are

27


not attributable to issues of capacity per se but rather the correct deployment of executive functions

(i.e., inhibition of distractors; Enright & Beech, 1993; Enright, Beech, & Claridge, 1995) within WM.

The recall of verbal information in the simple and complex domains revealed an interesting

pattern. First, across eight tasks, effect sizes ranged from large (d=0.97 for recall of complex verbal

information), medium (d=0.65 for WLM – story recall), and none (d=0.02 for the recall of neutral

words). This indicates that verbal memory performance in OCD is dependent on the task. In addition,

within the category of simple recall of verbal information, OCD participants suffered from a medium

impairment for threat words (d=0.46) compared to small and negligible impairments for cued

(d=0.22) and neutral (d=0.02) word recall, respectively. This pattern again indicates that WM is

generally intact in OCD, yet in these tasks, the simple introduction of a threatening word interferes

with the accuracy of its maintenance in WM to the detriment of subsequent memory recall. Again, in

the domain of recall of complex verbal information, we observe the largest impairments for tasks

(i.e., combined verbal tasks, WLM – story recall) which place a premium on executive function (e.g.,

semantic processing) as observed in the previous moderator analysis.

In the recognition domain, we observed that those with OCD had medium deficits for visuo-

spatial (d=0.57) compared to small impairments for object (d=0.27) and verbal (d=0.20) recognition.

However, as executive function did not moderate the effects of memory performance for those with

OCD in this domain, we cannot conclude specifically on how executive function contributes to each

of these tasks other than to say the findings matched those observed previously (Savage, Keuthen,

Jenike, Brown, Baer, Kendrick, Miguel, Rauch, & Albert, 1996).

To our knowledge, this is the first time that declarative and implicit memory domains have

been subject to meta-analytic review. For those with OCD, prospective memory (i.e., remember to

perform an action) showed a moderate impairment (d=0.43) relative to controls. In an experimental

study, Yang, Peng, Wang, Geng, Miao, Shum, Cheung, and Chan (2015) proposed that deficits in

prospective memory in OCD may be attributable to impairments in executive functions like updating

and mental-shifting. Snyder et al. (2015) identified updating as significantly impaired in OCD,

suggesting that such prospective memory impairments are secondary to executive dysfunction.

Source, false and procedural/motor memory resulted in small to no memory impairments (d=0.30,

0.10, -0.04, respectively). Interestingly, Shahar et al. (2017) reported that while procedural motor

memory was intact in OCD (d=0.10), they did observe ‘compensatory’ impairments in their ability to

classify/identify the stimulus. We propose that if procedural memory tasks were to employ

stimuli/complexity that tax executive function and binding complexity, then we may observe

impairments for those with OCD.

Limitations and Future Research

The present review validates the EBL system to understand memory performance in OCD.

However, we identify limitations within the present study, and where appropriate propose potential

28


solutions via the avenue of future research. First, due to a lack of studies with OCD-subtypes (e.g., 13

studies in the washer versus checker meta-analysis of Leopold & Backenstrass, 2015), and

comorbidities, it was not possible to conduct moderator analyses that compared, for example,

checkers versus cleaners or the impact of depression and/or anxiety on memory impairment in OCD.

Another issue in this meta-analysis and others is differences in how to categorise a given task. For

example, Shin et al. (2014) defined the digit span task as a measure of attention, whereas using our

inclusion criteria (i.e., the memorization of items a given sequence) we categorised it within the span-

sequence domain. We argue that as we classified tasks based on defined criterions, we have some

confidence on the internal validity of how we categorised tasks (see Shin, et al., 2014). In light of

this, we suggest that there is a need to provide reliable and valid task-scoring frameworks that future

meta-analysis and experimental studies can employ. This would create a body of literature that

characterises tasks in a congruent manner, which in turn would improve the ability to compare the

outcomes of different studies and to conduct meta-analysis in any given area.

A further shortcoming of many studies that became evident throughout the preliminary

search and coding of studies was that very few employed ecologically valid stimuli (e.g., Tolin,

Abramowitz, Brigidi, Amir, Street, & Foa, 2001; Harkin, et al., 2011). As such, while we observed an

informative pattern of memory performance for those with OCD, we are limited to the extent that we

conclude on memory performance for idiographic stimuli in OCD relevant settings. Indeed, some

studies that have utilized ecologically valid stimuli have reported memory biases in favour of threat-

relevant stimuli in OCD (Constans, Foa, Franklin, & Mathews, 1995; Radomsky & Rachman, 2004).

In future, if more studies employ ecologically valid stimuli, then a subsequent meta-analysis

comparing ecological to traditional stimuli (e.g., RCFT) in the context of the EBL taxonomy would

be informative. In addition, the extent to which established features of OCD, e.g., cognitive (e.g.,

intolerance of uncertainty; Tolin, Abramowitz, Brigidi, & Foa, 2003), meta-cognitive (e.g.,

confidence in memory; Tolin, et al., 2001), attitudinal (e.g., inflated personal responsibility;

Salkovskis, Wroe, Gledhill, Morrison, Forrester, Richards, Reynolds, & Thorpe, 2000) and/or

emotional (Thorsen, Hagland, Radua, Mataix-Cols, Kvale, Hansen, & van den Heuvel, 2018)

interacted with individual or collective dimensions of the EBL system is unknown and requires

investigation in future research.

Lastly, we propose the following to counter the inherent weakness of inferring cognitive

processes from outcome measures (see Ouimet, et al., 2019 discussed above). First, a possible

solution is to infer causality via the manipulation of dimensions of the EBL in non-OCD participants

(see van den Hout, van Dis, van Woudenberg, & van de Groep, 2019 for a review on such methods),

and measure changes in memory performance, memory confidence, and obsessional-compulsive

symptoms. Second, neuroimaging can inform the cognitive-emotional processes that contribute to

task performance, even when outcome measures are uninformative. For example, Henseler et al.

29


(2008) reported no differences in the performance of OCD patients and controls on simple tests of

WM. Based on this behavioural finding alone, one could conclude that as there is no difference in the

outcome measures, then there is no difference in the cognitive processes of these two groups.

However, Henseler and colleagues also conducted brain imaging and reported that those with OCD

had a greater activation in brain regions associated (i.e., compensatory processes) with basic rehearsal

and maintenance. Applying this to the EBL, using functional imaging in a synchronous manner with

specific manipulations in EBL dimensions, could inform the literature if memory performance in

OCD is attributable to processes of: WM maintenance (e.g., dorsolateral PFC; Ranganath, Cohen, &

Brozinsky, 2005), executive function/organizational strategies (e.g., orbitofrontal cortext; Choi,

Kang, Kim, Ha, Lee, Youn, Kim, Kim, & Kwon, 2004), binding (e.g., prefrontal cortex; Prabhakaran,

Narayanan, Zhao, & Gabrieli, 2000), load (e.g., frontoparietal network; Tomasi, Chang, Caparelli, &

Ernst, 2007), error monitoring/task complexity (e.g., anterior cingulate cortex; Koch, Wagner,

Schachtzabel, Peikert, Schultz, Sauer, & Schlosser, 2012) or emotional decision making (e.g.,

striatum; Crittenden, Tillberg, Riad, Shima, Gerfen, Curry, Housman, Nelson, Boyden, & Graybiel,

2016).

Conclusion

The present three-level meta-analysis of 305 effect sizes from 144 studies indicates that the

EBL taxonomy (Harkin & Kessler, 2011a) has explanatory power in explaining the memory

performance of those with OCD. Specifically, executive function appears to be the driving

mechanism behind the EBL framework’s predictive power for OCD memory performance, and

tellingly, negated effect size differences between visual and verbal tasks in those with OCD, when

executive demands were controlled. This highlights that it is the executive demands of a task and not

the visual or verbal description of a task, which determines memory performance in OCD. Domain-

specific moderator analyses and comparison of sub-task effect sizes were also generally in accord

with the cognitive parameters of the EBL taxonomy. We conclude that our novel approaches to

coding tasks along individual cognitive dimensions and the use of multi-level statistical analyses

provides a standardised means to examine multi-dimensional models of memory and cognitive

performance in OCD and other disorders.

30


References

References marked with an asterisk (*) indicate studies included in the meta-analysis.

Abramovitch, A., Abramowitz, J. S., & Mittelman, A. (2013). The neuropsychology of adult Obsessive-Compulsive Disorder: A meta-analysis. Clinical Psychology Review, 33(8), 1163-1171. doi: 10.1016/j.cpr.2013.09.004

Abramovitch, A., & Cooperman, A. (2015). The cognitive neuropsychology of Obsessive-Compulsive Disorder: A critical review. Journal of Obsessive-Compulsive and Related Disorders, 5, 24-36. doi: 10.1016/j.jocrd.2015.01.002

Adele, D. (2013). Executive functions. Annual Review of Psychology, 64, 135-168. doi: 10.1146/annurev-psych-113011-143750

Allen, R. J., Baddeley, A. D., & Hitch, G. J. (2006). Is the binding of visual features in working memory resource-demanding? Journal of Experimental Psychology-General, 135(2), 298-313. doi: 10.1037/0096-3445.135.2.298

Assink, M., & Wibbelink, C. J. M. (2016). Fitting three-level meta-analytic models in R: A step-by-step tutorial. The Quantitative Methods for Psychology, 12, 154-174.

Baddeley, A. (1986). Working Memory. Oxford, England: Clarendon Press.Baddeley, A., Allen, R. J., & Hitch, G. J. (2010). Investigating the episodic buffer. Psychologica

Belgica, 50, 223-243. doi: 10.5334/pb-50-3-4-223Baddeley, A. (2000). The episodic buffer: A new component of working memory? Trends in

Cognitive Sciences, 4(11), 417-423. doi: 10.1016/s1364-6613(00)01538-2Baddeley, A., Vallar, G., & Wilson, B. (1987). Sentence comprehension and phonological memory:

Some neuropsychological evidence. Attention and Performance, 12, 509-529. doi: 10.3758/BF03202730

Bohne, A., Savage, C. R., Deckersbach, T., Keuthen, N. J., Jenike, M. A., Tuschen-Caffier, B., & Wilhelm, S. (2005). Visuospatial abilities, memory, and executive functioning in trichotillomania and obsessive-compulsive disorder. Journal of Clinical and Experimental Neuropsychology, 27(4), 385-399. doi: 10.1080/13803390490520418

*Bohne, A., Keuthen, N. J., Tuschen-Caffier, B., & Wilhelm, S. (2005). Cognitive inhibition in trichotillomania and Obsessive-Compulsive Disorder. Behaviour Research and Therapy, 43(7), 923-942. doi: 10.1016/j.brat.2004.06.014

Boldrini, M., Del Pace, L., Placidi, G. P., Keilp, J., Ellis, S. P., Signori, S., Placidi, G. F., & Cappa, S.F. (2005). Selective cognitive deficits in obsessive-compulsive disorder compared to panic disorder with agoraphobia. Acta Psychiatrica Scandinavica, 111(2), 150-158. doi: 10.1111/j.1600-0447.2004.00247.x

Boone, K. B., Ananth, J., Philpott, L., Kaur, A., & Djenderedjian, A. (1991). Neuropsychological characteristics of nondepressed adults with Obsessive-Compulsive Disorder. Cognitive and Behavioral Neurology, 4(2), 96-109.

*Borges, M. C., Braga, D. T., Iego, S., D'Alcante, C. C., Sidrim, I., Machado, M. C., Pinto, P. S., Cordioli, A. V., do Rosario, M. C., Petribu, K., Mendlowicz, M. V., Mari, J. J., Miguel, E. C.,& Fontenelle, L. F. (2011). Cognitive dysfunction in post-traumatic Obsessive-Compulsive Disorder. Australian and New Zealand Journal of Psychiatry, 45(1), 76-85. doi: 10.3109/00048674.2010.527822

Bragdon, L. B., Gibb, B. E., & Coles, M. E. (2018). Does neuropsychological performance in OCD relate to different symptoms? A meta-analysis comparing the symmetry and obsessing dimensions. Depression and Anxiety, 35(8), 761-774. doi: 10.1002/da.22785

Brown, L. A. (2016). Spatial-sequential working memory in younger and older adults: Age predicts backward recall performance within both age groups. [Original Research]. Frontiers in Psychology, 7(1514). doi: 10.3389/fpsyg.2016.01514

Cabrera, A. R., McNally, R. J., & Savage, C. R. (2001). Missing the forest for the trees? Deficient memory for linguistic gist in obsessive-compulsive disorder. Psychological Medicine, 31(6), 1089-1094. doi: 10.1017/s0033291701004354

31


Cha, K. R., Koo, M. S., Kim, C. H., Kim, J. W., Oh, W. J., Suh, H. S., & Lee, H. S. (2008). Nonverbal memory dysfunction in obsessive-compulsive disorder patients with checking compulsions. Depression and Anxiety, 25(11), E115-120. doi: 10.1002/da.20377

Cheung, M. W. L. (2014). Modeling dependent effect sizes with three-level meta-analyses: A structural equation modeling approach. Psychological Methods, 19(2), 211-229. doi: 10.1037/a0032968

Chlebowski, C. (2011). Wechsler Memory Scale All Versions. In J. S. Kreutzer, J. DeLuca & B. Caplan (Eds.), Encyclopedia of Clinical Neuropsychology (pp. 2688-2690). New York, NY: Springer New York.

*Choi, J. S., Kang, D. H., Kim, J. J., Ha, T. H., Lee, J. M., Youn, T., Kim, I. Y., Kim, S. I., & Kwon, J. S. (2004). Left anterior subregion of orbitofrontal cortex volume reduction and impaired organizational strategies in Obsessive-Compulsive Disorder. Journal of Psychiatric Research, 38(2), 193-199. doi: 10.1016/j.jpsychires.2003.08.001

Christensen, K. J., Kim, S. W., Dysken, M. W., & Hoover, K. M. (1992). Neuropsychological performance in Obsessive–Compulsive Disorder. Biological Psychiatry, 31(1), 4-18. doi: 10.1016/0006-3223(92)90003-I

Ciesielski, K. T., Hamalainen, M. S., Geller, D. A., Wilhelm, S., Goldsmith, T. E., & Ahlfors, S. P. (2007). Dissociation between MEG alpha modulation and performance accuracy on visual working memory task in obsessive compulsive disorder. Human Brain Mapping, 28(12), 1401-1414. doi: 10.1002/hbm.20365

Ciesielski, K. T., Hamalainen, M. S., Lesnik, P. G., Geller, D. A., & Ahlfors, S. P. (2005). Increased MEG activation in OCD reflects a compensatory mechanism specific to the phase of a visual working memory task. Neuroimage, 24(4), 1180-1191. doi: 10.1016/j.neuroimage.2004.10.018

Clayton, I. C., Richards, J. C., & Edwards, C. J. (1999). Selective attention in Obsessive-Compulsive Disorder. Journal of Abnormal Psychology, 108(1), 171-175. doi: 10.1037//0021-843x.108.1.171

Coetzer, B. R. (2004). Obsessive-Compulsive Disorder following brain injury: A review. The International Journal of Psychiatry in Medicine, 34(4), 363-377. doi: 10.2190/xenn-nnwt-7n2k-r26a

Cohen, Y., Lachenmeyer, J. R., & Springer, C. (2003). Anxiety and selective attention in obsessive-compulsive disorder. [Article]. Behaviour Research and Therapy, 41(11), 1311-1323. doi: 10.1016/s0005-7967(03)00037-8

Coles, M. E., & Heimberg, R. G. (2002). Memory biases in the anxiety disorders: Current status. [Article]. Clinical Psychology Review, 22(4), 587-627. doi: 10.1016/s0272-7358(01)00113-1

Collette, F., Van der Linden, M., & Ponceret, M. (2000). Working memory, long-term memory, and language processing: Issues and future directions. Brain and Language, 71(1), 46-51. doi: 10.1006/brln.1999.2209

Constans, J. I., Foa, E. B., Franklin, M. E., & Mathews, A. (1995). Memory for actual and imagined events in OC checkers. Behaviour Research and Therapy, 33(6), 665-671. doi: 10.1016/0005-7967(94)00095-2

Corsi, P. M. (1972). Human memory and the medial temporal region of the brain: McGill University.Cowan, N. (1999). An embedded-process model of working memory. In A. Miyake & P. Shah.

(Eds.), Models of working memory: Mechanisms of active maintenance and executive control(pp. 62-101). New York; NY: Cambridge University Press.

Crittenden, J. R., Tillberg, P. W., Riad, M. H., Shima, Y., Gerfen, C. R., Curry, J., Housman, D. E., Nelson, S. B., Boyden, E. S., & Graybiel, A. M. (2016). Striosome–dendron bouquets highlight a unique striatonigral circuit targeting dopamine-containing neurons. Proceedings of the National Academy of Sciences, 201613337. doi: 10.1073/pnas.1613337113

de Fockert, J. W., Rees, G., Frith, C. D., & Lavie, N. (2001). The Role of Working Memory in VisualSelective Attention. Science, 291(5509), 1803-1806. doi: 10.1126/science.1056496

Deckersbach, T., Otto, M. W., Savage, C. R., Baer, L., & Jenike, M. A. (2000). The relationship between semantic organization and memory in obsessive-compulsive disorder. Psychotherapy and Psychosomatics, 69(2), 101-107. doi: 10.1159/000012373

32


Del Casale, A., Rapinesi, C., Kotzalidis, G. D., De Rossi, P., Curto, M., Janiri, D., Criscuolo, S., Alessi, M. C., Ferri, V. R., De Giorgi, R., Sani, G., Ferracuti, S., Girardi, P., & Brugnoli, R. (2016). Executive functions in obsessive-compulsive disorder: An activation likelihood estimate meta-analysis of fMRI studies. World Journal of Biological Psychiatry, 17(5), 378-393. doi: 10.3109/15622975.2015.1102323

Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (1988). The California Verbal Learning Test. San Antonio, Texas.: The Psychological Corporation.

Duval, S., & Tweedie, R. (2000). Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics, 56, 455-463. doi: 10.1111/j.0006-341X.200000455.x

Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ, 315(7109), 629-634. doi: 10.1136/bmj.315.7109.629

Enright, S. J., & Beech, A. R. (1993). Reduced cognitive inhibition in obsessive-compulsive disorder.British Journal of Clinical Psychology, 32(1), 67-74. doi: 10.1111/j.2044-8260.1993.tb01028.x

Enright, S. J., Beech, A. R., & Claridge, G. S. (1995). A further investigation of cognitive inhibition in Obsessive-Compulsive Disorder and other anxiety disorders. Personality and Individual Differences, 19(4), 535-542.

Exner, C., Martin, V., & Rief, W. (2009). Self-focused ruminations and memory deficits in Obsessive-Compulsive Disorder. Cognitive Therapy and Research, 33(2), 163-174. doi: 10.1007/s10608-007-9162-x

Exner, C., Kohl, A., Zaudig, M., Langs, G., Lincoln, T. M., & Rief, W. (2009). Metacognition and episodic memory in obsessive-compulsive disorder. Journal of Anxiety Disorders, 23(5), 624-631. doi: 10.1016/j.janxdis.2009.01.010

Fougnie, D., & Marois, R. (2009). Attentive tracking disrupts feature binding in visual working memory. Visual Cognition, 17, 48-66. doi: 10.1080/13506280802281337

Fradkin, I., Strauss, A. Y., Pereg, M., & Huppert, J. D. (2018). Rigidly applied rules? Revisiting inflexibility in Obsessive Compulsive Disorder using multilevel meta-analysis. Clinical Psychological Science, 6(4), 481-505. doi: 10.1177/2167702618756069

Friedman, N. P., Miyake, A., Corley, R. P., Young, S. E., DeFries, J. C., & Hewitt, J. K. (2006). Not all executive functions are related to intelligence. Psychological Science, 17(2), 172-179. doi:10.1111/j.1467-9280.2006.01681.x

Gao, Z., Wu, F., Qiu, F., He, K., Yang, Y., & Shen, M. (2017). Bindings in working memory: The role of object-based attention. Attention, Perception, & Psychophysics, 79(2), 533-552. doi: 10.3758/s13414-016-1227-z

Greisberg, S., & McKay, D. (2003). Neuropsychology of obsessive-compulsive disorder: A review and treatment implications. Clinical Psychology Review, 23(1), 95-117. doi: 10.1016/s0272-7358(02)00232-5

*Grisham, J. R., & Williams, A. D. (2013). Responding to intrusions in obsessive-compulsive disorder: The roles of neuropsychological functioning and beliefs about thoughts. Journal of Behavior Therapy and Experimental Psychiatry, 44(3), 343-350. doi: 10.1016/j.jbtep.2013.01.005

Harkin, B., & Kessler, K. (2009). How checking breeds doubt: Reduced performance in a simple working memory task. Behaviour Research and Therapy, 47(6), 504-512. doi: 10.1016/j.brat.2009.03.002.

Harkin, B., & Kessler, K. (2012a). Deficient Inhibition-Of-Return in subclinical OCD when attention is directed to the threatening aspects of a stimulus. Depression and Anxiety, 29(9), 807-815. doi: 10.1002/da.21911

Harkin, B., & Kessler, K. (2011a). The role of working memory in compulsive checking and OCD: Asystematic classification of 58 experimental findings. Clinical Psychology Review, 31(6), 1004-1021. doi: 10.1016/j.cpr.2011.06.004.

*Harkin, B., Rutherford, H., & Kessler, K. (2011). Impaired executive functioning in subclinical compulsive checking with ecologically valid stimuli in a working memory task. Frontiers in Psychology, 2, 78. doi: 10.3389/fpsyg.2011.00078

33


Harkin, B., & Kessler, K. (2011b). How checking as a cognitive style influences working memory performance. Applied Cognitive Psychology, 25(2), 219-228. doi: 10.1002/acp.1666

Harkin, B., Miellet, S., & Kessler, K. (2012b). What checkers actually check: An eye tracking study of working memory and executive control. PloS One, 7(9). doi: 10.1371/journal.pone.0044689

Harrer, M., Cuijpers, P., & Ebert, D. D. (2019). Doing Meta-Analysis in R: A Hands-on Guide. https://bookdown.org/MathiasHarrer/Doing\_ Meta\_ Analysis\_ in\_ R/.

Head, D., Bolton, D., & Hymas, N. (1989). Deficit in cognitive shifting ability in patients with obsessive-compulsive disorder. Biological Psychiatry, 25(7), 929-937. doi: 10.1016/0006-3223(89)90272-2

Henseler, I., Gruber, O., Kraft, S., Krick, C., Reith, W., & Falkai, P. (2008). Compensatory hyperactivations as markers of latent working memory dysfunctions in patients with obsessive-compulsive disorder: An fMRI study. Journal of Psychiatry and Neuroscience, 33(3), 209-215. doi: PMID18592040

Hermans, D., Engelen, U., Grouwels, L., Joos, E., Lemmens, J., & Pieters, G. (2008). Cognitive confidence in obsessive-compulsive disorder: Distrusting perception, attention and memory. [Article]. Behaviour Research and Therapy, 46(1), 98-113. doi: 10.1016/j.brat.2007.11.001

Hinton, G. E., McClelland, J. L., & Rumelhart, D. E. (1986). Distributed representations. In D. E. Rumelhart & J. L. McClelland (Eds.), Parallel distributed processing: Explorations in the microstructure of cognition (Vol. 1, pp. 77–109). Cambridge: MIT Press.

Hox, J. J. (2010). Multilevel analysis: Techniques and applications (2nd ed ed.): New York, NY, US: Routledge/Taylor & Francis Group.

Hyun, J. S., Woodman, G. F., & Luck, S. J. (2009). The role of attention in the binding of surface features to locations. Visual Cognition, 17(1-2), 10-24. doi: 10.1080/13506280802113894

Irak, M., & Flament, M. F. (2009). Attention in sub-clinical obsessive-compulsive checkers. Journal of Anxiety Disorders, 23(3), 320-326. doi: 10.1016/j.janxdis.2009.01.007

*Jelinek, L., Rietschel, L., Kellner, M., Muhtz, C., & Moritz, S. (2012). The effect of practice on the recall of salient information in obsessive-compulsive disorder. Psychiatry Research, 198(1), 89-93. doi: 10.1016/j.psychres.2012.03.002

Jelinek, L., Moritz, S., Heeren, D., & Naber, D. (2006). Everyday memory functioning in obsessive- compulsive disorder. Journal of the International Neuropsychological Society, 12(5), 746-749. doi: 10.1017/S1355617706060899

Kane, M. J., & Engle, R. W. (2002). The role of prefrontal cortex in working-memory capacity, executive attention, and general fluid intelligence: An individual-differences perspective. Psychonomic Bulletin & Review, 9(4), 637-671. doi: 10.3758/BF03196323

Kirchner, W. K. (1958). Age differences in short-term retention of rapidly changing information. Journal of Experimental Psychology, 55(4), 352-358. doi: 10.1037/h0043688

Koch, K., Wagner, G., Schachtzabel, C., Peikert, G., Schultz, C. C., Sauer, H., & Schlosser, R. G. (2012). Aberrant anterior cingulate activation in obsessive-compulsive disorder is related to task complexity. Neuropsychologia, 50(5), 958-964. doi: 10.1016/j.neuropsychologia.2012.02.002

Kuelz, A. K., Hohagen, F., & Voderholzer, U. (2004). Neuropsychological performance in obsessive-compulsive disorder: A critical review. Biological Psychology, 65(3), 185-236. doi: 10.1016/j.biopsycho.2003.07.007

Langerock, N., Vergauwe, E., & Barrouillet, P. (2014). The maintenance of cross-domain associations in the episodic buffer. Journal of Experimental Psychology-Learning Memory and Cognition, 40(4), 1096-1109. doi: 10.1037/a0035783

Leopold, R., & Backenstrass, M. (2015). Neuropsychological differences between obsessive-compulsive washers and checkers: A systematic review and meta-analysis. Journal of Anxiety Disorders, 30, 48-58. doi: 10.1016/j.janxdis.2014.12.016

Lezak, M. D. (1995). Neuropsychological Assessment (3rd ed.): New York: Oxofrd University Press.Lezak, M. D., Howieson, D. B., & Loring, D. W. (2004). Neuropsychological Assessment. New

York: Oxford University Press.Luck, S. J., & Vogel, E. K. (1997). The capacity of visual working memory for features and

conjunctions. Nature, 390(6657), 279-281. doi: 10.1038/36846

34

https://bookdown.org/MathiasHarrer/Doing


MacDonald, P. A., Antony, M. M., Macleod, C. M., & Richter, M. A. (1997). Memory and confidence in memory judgements among individuals with obsessive compulsive disorder and non-clinical controls. Behaviour Research and Therapy, 35(6), 497-505. doi: 10.1016/s0005-7967(97)00013-2

Martin, A., Wiggs, C. L., Altemus, M., Rubenstein, C., & Murphy, D. L. (1995). Working-memory asassessed by subject-ordered tasks in patients with Obsessive-Compulsive Disorder. Journal of Clinical and Experimental Neuropsychology, 17(5), 786-792. doi: 10.1080/01688639508405167

*Mataix-Cols, D. (2003). Declarative and procedural learning in individuals with subclinical obsessive-compulsive symptoms. Journal of Clinical and Experimental Neuropsychology, 25(6), 830-841. doi: 10.1076/jcen.25.6.830.16477

McDaniel, M. A., & Einstein, G. O. (2007). Prospective memory: An overview and synthesis of an emerging field. Sage Publications Ltd.

*McNally, R. J., & Kohlbeck, P. A. (1993). Reality monitoring in obsessive-compulsive disorder. Behaviour Research and Therapy, 31(3), 249-253. doi: 10.1016/0005-7967(93)90023-n

Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63(81-97). doi: 10.1037/h0043158

Miyake, A., & Shah, P. (1999). Models of working memory: Mechanisms of active maintenance and executive control. New York, NY: Cambridge University Press.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their contributions to complex "frontal lobe" tasks: A latent variable analysis. Cognitive Psychology, 41(1), 49-100. doi: 10.1006/cogp.1999.0734

Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematicreviews and meta-analyses: the PRISMA statement. BMJ, 339, b2535. doi: 10.1136/bmj.b2535

Moreno-Castilla, P., Guzman-Ramos, K., & Bermudez-Rattoni, F. (2018). Chapter 28 - Object Recognition and Object Location Recognition Memory – The Role of Dopamine and Noradrenaline Handbook of Behavioral Neuroscience (Vol. 27, pp. 403-413).

Morey, C. C. (2009). Integrated cross-domain object storage in working memory: Evidence from a verbal-spatial memory task. Quarterly Journal of Experimental Psychology, 62(11), 2235-2251. doi: 10.1080/17470210902763382

Muller, J., & Roberts, J. E. (2005). Memory and attention in Obsessive-Compulsive Disorder: A review. Journal of Anxiety Disorders, 19(1), 1-28. doi: 10.1016/j.janxdis.2003.12.001

Olley, A., Malhi, G., & Sachdev, P. (2007). Memory and executive functioning in obsessive-compulsive disorder: A selective review. Journal of Affective Disorders, 104(1-3), 15-23. doi: 10.1016/j.jad.2007.02.023

Omori, I. M., Murata, Y., Yamanishi, T., Nakaaki, S., Akechi, T., Mikuni, M., & Furukawa, T. A. (2007). The differential impact of executive attention dysfunction on episodic memory in obsessive-compulsive disorder patients with checking symptoms vs. those with washing symptoms. Journal of Psychiatric Research, 41(9), 776-784. doi: 10.1016/j.jpsychires.2006.05.005

Ouimet, A. J., Ashbaugh, A. R., & Radomsky, A. S. (2019). Hoping for more: How cognitive sciencehas and hasn't been helpful to the OCD clinician. Clinical Psychology Review, 69, 14-29. doi:10.1016/j.cpr.2018.04.003

Ouzzani, M., Hammady, H., Fedorowicz, Z., & Elmagarmid, A. (2016). Rayyan-a web and mobile app for systematic reviews. Systematic Reviews, 5, 210.

Pandey, J. (2011). Source Memory. In J. S. Kreutzer, J. DeLuca & B. Caplan (Eds.), Encyclopedia of Clinical Neuropsychology (pp. 2325-2326). New York, NY: Springer New York.

Parr, W. V. (1992). Delayed matching-to-sample performance as a measure of human visuospatial working memory. Bulletin of the Psychonomic Society, 30(5), 369-372. doi: 10.3758/BF03334092

Penades, R., Catalan, R., Andres, S., Salamero, M., & Gasto, C. (2005). Executive function and nonverbal memory in obsessive-compulsive disorder. Psychiatry Research, 133(1), 81-90. doi: 10.1016/j.psychres.2004.09.005

35


Peters, J. L., Sutton, A. J., Jones, D. R., Abrams, K. R., & Rushton, L. (2007). Performance of the trim and fill method in the presence of publication bias and between-study heterogeneity. Statistics in Medicine, 26(25), 4544-4562. doi: 10.1002/sim.2889

Petrides, M., & Milner, B. (1982). Deficits on subjectordered tasks after frontal- and temporal-lobe lesions in man. Neuropsychologia, 20(3), 249-262. doi: 10.1016/0028-3932(82)90100-2

Phillips, L. H., Wynn, V., Gilhooly, K. J., Della Sala, S., & Logie, R. H. (1999). The role of memory in the Tower of London task. Memory, 7, 209-232. doi: 10.1080/741944066

Prabhakaran, V., Narayanan, K., Zhao, Z., & Gabrieli, J. D. E. (2000). Integration of diverse information in working memory within the frontal lobe. [Article]. Nature Neuroscience, 3(1),85-90.

Radomsky, A. S., & Rachman, S. (2004). The importance of importance in OCD memory research. Journal of Behavior Therapy and Experimental Psychiatry, 35(2), 137-151. doi: 10.1016/j.jbtep.2004.04.005

Ranganath, C., Cohen, M. X., & Brozinsky, C. J. (2005). Working memory maintenance contributes to long-term memory formation: neural and behavioral evidence. [Review]. Journal of Cognitive Neuroscience, 17(7), 994-1010. doi: 10.1162/0898929054475118

Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 803-814. doi: 10.1037/0278-7393.21.4.803

Rojdev, R. S., Krikorian, R., Feldman, D., Tracy, A., & Williams, C. (1998). Working memory and concept formation on the Tower of London task. Archives of Clinical Neuropsychology, 13, 32-33. doi: 10.1093/arclin/13.1.32a

Rosenthal, R. (1986). Meta-Analytic Procedures for Social Science Research (Vol. 15). Beverly Hills: Sage Publications.

*Roth, R. M., Baribeau, J., Milovan, D., O'Connor, K., & Todorov, C. (2004). Procedural and declarative memory in obsessive-compulsive disorder. Journal of the International Neuropsychological Society, 10(5), 647-654. doi: 10.1017/S1355617704105018

Salkovskis, P. M., Wroe, A. L., Gledhill, A., Morrison, N., Forrester, E., Richards, C., Reynolds, M., & Thorpe, S. (2000). Responsibility attitudes and interpretations are characteristic of obsessive compulsive disorder. Behaviour Research and Therapy, 38(4), 347-372. doi: 10.1016/s0005-7967(99)00071-6

*Savage, C. R., Deckersbach, T., Wilhelm, S., Rauch, S. L., Baer, L., Reid, T., & Jenike, M. A. (2000). Strategic processing and episodic memory impairment in obsessive compulsive disorder. Neuropsychology, 14(1), 141-151. doi: 10.1037//0894-4105.14.1.141

Savage, C. R., Keuthen, N. J., Jenike, M. A., Brown, H. D., Baer, L., Kendrick, A. D., Miguel, E. C., Rauch, S. L., & Albert, M. S. (1996). Recall and recognition memory in obsessive-compulsive disorder. Journal of Neuropsychiatry and Clinical Neurosciences, 8(1), 99-103. doi: 10.1176/jnp.8.1.99

*Savage, C. R., Baer, L., Keuthen, N. J., Brown, H. D., Rauch, S. L., & Jenike, M. A. (1999). Organizational strategies mediate nonverbal memory impairment in obsessive-compulsive disorder. Biological Psychiatry, 45(7), 905-916. doi: 10.1016/s0006-3223(98)00278-9

Shahar, N., Teodorescu, A. R., Anholt, G. E., Karmon-Presser, A., & Meiran, N. (2017). Examining procedural working memory processing in obsessive-compulsive disorder. Psychiatry Research, 253, 197-204. doi: 10.1016/j.psychres.2017.03.048

Shallice, T., 1982. Specific impairments of planning. (1982). Specific impairments of planning. Philosophical Transactions of The Royal Society of London Series B, 298, 199–209. doi: 10.1098/rstb.1982.0082

Shin, N. Y., Lee, T. Y., Kim, E., & Kwon, J. S. (2014). Cognitive functioning in obsessive-compulsive disorder: A meta-analysis. Psychological Medicine, 44(6), 1121-1130. doi: 10.1017/S0033291713001803

Simon, S. S., Tusch, E. S., Holcomb, P. J., & Daffner, K. R. (2016). Increasing working memory loadreduces processing of cross-modal task-irrelevant stimuli even after controlling for task difficulty and executive capacity. [Original Research]. Frontiers in Human Neuroscience, 10(380). doi: 10.3389/fnhum.2016.00380

36


Simonsohn, U. (Producer). (2017). The funnel plot is invalid because of this crazy assumption: r(n,d)=0. Thinking about evidence, and vice versa.

Smith, E. E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science, 283(5408), 1657-1661. doi: 10.1126/science.283.5408.1657

Snyder, H. R., Kaiser, R. H., Warren, S. L., & Heller, W. (2015). Obsessive-compulsive disorder is associated with broad impairments in executive function: A meta-analysis. Clinical Psychological Science, 3(2), 301-330. doi: 10.1177/2167702614534210

Sörqvist, P., & Rönnberg, J. (2014). Individual differences in distractibility: An update and a model. PsyCh Journal, 3(1), 42-57. doi: 10.1002/pchj.47

Sternberg, S. (1966). High-speed scanning in human memory. Science, 153(3736), 652-654. doi: 10.1126/science.153.3736.652

Stratta, P., Daneluzzo, E., Prosperini, P., Bustini, M., Mattei, P., & Rossi, A. (1997). Is wisconsin card sorting test performance related to ‘working memory’ capacity? Schizophrenia Research, 27(1), 11-19. doi: https://doi.org/10.1016/S0920-9964(97)00090-X

Tallis, F. (1997). The neuropsychology of obsessive-compulsive disorder: A review and consideration of clinical implications. British Journal of Clinical Psychology, 36(1), 3-20. doi: 10.1111/j.2044-8260.1997.tb01226.x

Team, R. C. (Producer). (2013). R: A language and environment for statistical computing. Thalmann, M., Souza, A. S., & Oberauer, K. (2019). How does chunking help working memory?

Journal of Experimental Psychology: Learning, Memory, and Cognition, 45(1), 37-55. doi: 10.1037/xlm0000578

Thorsen, A. L., Hagland, P., Radua, J., Mataix-Cols, D., Kvale, G., Hansen, B., & van den Heuvel, O.A. (2018). Emotional processing in Obsessive-Compulsive Disorder: A systematic review and meta-analysis of 25 functional neuroimaging studies. Biological Psychiatry-Cognitive Neuroscience and Neuroimaging, 3(6), 563-571. doi: 10.1016/j.bpsc.2018.01.009

Tolin, D. F., Villavicencio, A., Umbach, A., & Kurtz, M. M. (2011). Neuropsychological functioning in hoarding disorder. Psychiatry Research, 189, 413-418. doi: 10.1016/j.psychres.2011.06.022

Tolin, D. F., Abramowitz, J. S., Brigidi, B. D., & Foa, E. B. (2003). Intolerance of uncertainty in obsessive-compulsive disorder. [Article]. Journal of Anxiety Disorders, 17(2), 233-242. doi: 10.1016/s0887-6185(02)00182-2

Tolin, D. F., Abramowitz, J. S., Brigidi, B. D., Amir, N., Street, G. P., & Foa, E. B. (2001). Memory and memory confidence in obsessive-compulsive disorder. Behaviour Research and Therapy,39(8), 913-927. doi: 10.1016/s0005-7967(00)00064-4

Tomasi, D., Chang, L., Caparelli, E. C., & Ernst, T. (2007). Different activation patterns for working memory load and visual attention load. Brain Research, 1132(1), 158-165. doi: 10.1016/j.brainres.2006.11.030

Treisman, A. (1996). The binding problem. Current Opinion in Neurobiology, 6(2), 171-178. doi: 10.1016/s0959-4388(96)80070-5

Treisman, A., & Zhang, W. (2006). Location and binding in visual working memory. Memory and Cognition, 34(8), 1704-1719. doi: 10.3758/bf03195932

Turner, M. L., & Engle, R. W. (1989). Is working memory capacity task dependent? Journal of Memory and Language, 28, 127-154. doi: 10.1016/0749-596X(89)90040-5

van den Hout, M. A., van Dis, E. A. M., van Woudenberg, C., & van de Groep, I. H. (2019). OCD-like checking in the lab: A meta-analysis and improvement of an experimental paradigm. Journal of Obsessive-Compulsive and Related Disorders, 20, 39-49. doi: 10.1016/j.jocrd.2017.11.006

Van den Noortgate, W., López-López, J. A., Marín-Martínez, F., & Sánchez-Meca, J. (2013). Three-level meta-analysis of dependent effect sizes. Behavior Research Methods, 45(2), 576-594. doi: 10.3758/s13428-012-0261-6

*van der Wee, N. J., Ramsey, N. F., Jansma, J. M., Denys, D. A., van Megen, H. J., Westenberg, H. M., & Kahn, R. S. (2003). Spatial working memory deficits in obsessive compulsive disorderare associated with excessive engagement of the medial frontal cortex. Neuroimage, 20(4), 2271-2280. doi: 10.1016/j.neuroimage.2003.05.001

37

https://doi.org/10.1016/S0920-9964(97)00090-X


van der Wee, N. J., Ramsey, N. F., van Megen, H. J., Denys, D., Westenberg, H. G., & Kahn, R. S. (2007). Spatial working memory in obsessive-compulsive disorder improves with clinical response: A functional MRI study. European Neuropsychopharmacology, 17(1), 16-23. doi: 10.1016/j.euroneuro.2006.04.012

VanRullen, R. (2009). Binding hardwired versus on-demand feature conjunctions. Visual Cognition, 17(1-2), 103-119.

Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal of Statistical Software, 36, 1-48.

Welsh, M. C., Satterlee-Cartmell, T., & Stine, M. (1999). Towers of Hanoi and London: Contributionof working memory and inhibition to performance. Brain and Cognition, 41, 231-242. doi: 10.1006/brcg.1999.1123

Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York.Wilson, D. B. (2019). Practical Meta-Analysis Effect Size Calculator: Find link at:

http://www.campbellcollaboration.org/escalc/html/EffectSizeCalculator-Home.php. Wolters, G., & Raffone, A. (2008). Coherence and recurrency: Maintenance, control and integration

in working memory. Cognitive Processing, 9(1), 1-17. doi: 10.1007/s10339-007-0185-8*Yang, T. X., Peng, Z. W., Wang, Y., Geng, F. L., Miao, G. D., Shum, D. H., Cheung, E. F., & Chan,

R. C. (2015). The nature of prospective memory deficit in patients with obsessive-compulsivedisorder. Psychiatry Research, 230(2), 479-486. doi: 10.1016/j.psychres.2015.09.041

Zielinski, C. M., Taylor, M. A., & Juzwin, K. R. (1991). Neuropsychological deficits in obsessive-compulsive disorder. Neuropsychiatry, Neuropsychology, and Behavioral Neurology, 4(2), 110-126.

38

http://www.campbellcollaboration.org/escalc/html/EffectSizeCalculator-Home.php

multi-dimension model of memory in ocd: a-meta analysis

Documents